JP2010230452A - Method for inspecting defect and defect inspection system - Google Patents

Method for inspecting defect and defect inspection system Download PDF

Info

Publication number
JP2010230452A
JP2010230452A JP2009077696A JP2009077696A JP2010230452A JP 2010230452 A JP2010230452 A JP 2010230452A JP 2009077696 A JP2009077696 A JP 2009077696A JP 2009077696 A JP2009077696 A JP 2009077696A JP 2010230452 A JP2010230452 A JP 2010230452A
Authority
JP
Japan
Prior art keywords
pixel
defective
setting
range
image
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
JP2009077696A
Other languages
Japanese (ja)
Inventor
Kensho Sugimoto
憲昭 杉本
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Panasonic Electric Works Co Ltd
Original Assignee
Panasonic Electric Works Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Panasonic Electric Works Co Ltd filed Critical Panasonic Electric Works Co Ltd
Priority to JP2009077696A priority Critical patent/JP2010230452A/en
Publication of JP2010230452A publication Critical patent/JP2010230452A/en
Withdrawn legal-status Critical Current

Links

Images

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of inspecting defects and a defect inspection system, capable of updating a conforming article range of light and shade values at all times for optimization by relatively simple processing. <P>SOLUTION: An image processing unit 1 includes a range-updating means 10 for adding a captured image of an object 4 to be inspected to an image memory 14 for setting as an image for setting, when a defect determining means 13 determines that the object is a conforming article. An arithmetic processing unit 15 not only sets a conforming article range from the image for setting stored in the image memory 14 for setting beforehand, but also obtains a frequency distribution of light and shade values for each pixel each time, when the range updating means 10 stores a new image for setting in the image memory 14 for setting, and resets the conforming article range from the result of the frequency distribution. However, each time the defect determining means 13 determines that the object is a conforming article, the range updating means 10 adds the image for setting to the image memory 14 for setting and automatically updates the conforming article range, based on a plurality of images for setting including the added image for setting. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

本発明は、同一仕様の多数の検査対象物を順次撮像して得られる画像を用いて、画像処理により各検査対象物の欠陥の有無を検査する欠陥検査方法および欠陥検査システムに関するものである。   The present invention relates to a defect inspection method and a defect inspection system for inspecting each inspection object for defects by image processing using images obtained by sequentially imaging a large number of inspection objects of the same specification.

たとえば電子部品材料、医薬品、食品などの製造に当たって、同一仕様の物品を量産する場合に、撮像された検査対象物の画像を検査員が目視で確認する方法では検査精度が悪く効率もよくないため、画像処理によって欠陥の有無を自動的に判定できる上記欠陥検査方法を用いることが多い。   For example, when mass-producing articles with the same specifications in the production of electronic component materials, pharmaceuticals, foods, etc., the method in which the inspector visually confirms the image of the imaged inspection object is not accurate and efficient. In many cases, the defect inspection method that can automatically determine the presence or absence of a defect by image processing is used.

一般的に、この種の欠陥検査に用いられる欠陥検査システムは、検査対象物に光またはX線を照射する照射装置と、検査対象物を撮像して濃淡値を画素値とした撮像画像を得る撮像装置と、撮像装置で得られた撮像画像から検査対象物の欠陥の有無を検査する画像処理装置とを備えている。すなわち、検査対象物に可視光を照射した状態で得られた撮像画像からは、検査対象物の外観上の欠陥(ひび割れ、キズ等)を検査することができる。また、検査対象物にX線を照射して得られた検査対象物のX線透過画像を用いれば、検査対象物を破壊せずに検査対象物の内部の欠陥(異物混入、気泡等)を検査することが可能である(たとえば特許文献1参照)。   In general, a defect inspection system used for this type of defect inspection obtains an irradiation device that irradiates an inspection target with light or X-rays, and a captured image in which the inspection target is imaged and gray values are used as pixel values. An imaging apparatus and an image processing apparatus that inspects the presence or absence of a defect of an inspection object from a captured image obtained by the imaging apparatus are provided. That is, defects (cracks, scratches, etc.) on the appearance of the inspection object can be inspected from the captured image obtained in a state where the inspection object is irradiated with visible light. Further, if an X-ray transmission image of the inspection object obtained by irradiating the inspection object with X-rays is used, defects (foreign matter contamination, bubbles, etc.) inside the inspection object can be obtained without destroying the inspection object. It is possible to inspect (for example, refer to Patent Document 1).

この種の欠陥検査方法として、予め欠陥のない良品を撮像した画像(設定用画像)によって画素ごとに濃淡値について良品となる範囲を規定しておき、検査対象物を撮像して得られた撮像画像の各画素の濃淡値を良品範囲と比較し、濃淡値が良品範囲にないと判定された画素数に応じて検査対象物の欠陥の有無を判定することが考えられる。この方法では、撮像画像において濃淡値が良品範囲外となる画素が規定数を超える場合に、当該検査対象物に欠陥があるものと判定される。   As this type of defect inspection method, an image obtained by imaging an inspection object by prescribing a range in which a gray value is determined for each pixel by an image (setting image) obtained by imaging a non-defective product in advance. It is conceivable to compare the gray value of each pixel of the image with the non-defective range and determine the presence or absence of a defect in the inspection object according to the number of pixels for which the gray value is determined not to be within the non-defective range. In this method, when the number of pixels whose grayscale value is outside the non-defective range exceeds the specified number in the captured image, it is determined that the inspection object is defective.

ところで、多数の検査対象物を順次検査するに当たっては、たとえば照射装置の出力低下などによって撮像画像の濃淡値が徐々に変化することがあるので、濃淡値について最初に設定された良品範囲が必ずしも最適な値とは限らず、濃淡値の良品範囲を随時更新して最適化することが望ましい。このように閾値を随時更新して最適化するための手法として、ニューラルネットワーク等の学習型アルゴリズムを用いることが考えられている。   By the way, when sequentially inspecting a large number of inspection objects, the gray value of the captured image may gradually change due to, for example, a decrease in the output of the irradiation device. It is not always a good value, but it is desirable to optimize the good and dark range of the gray value by updating it as needed. As a technique for updating and optimizing the threshold value as needed in this way, it is considered to use a learning type algorithm such as a neural network.

特開昭59−81544号公報JP 59-81544 A

しかし、ニューラルネットワーク等の学習型アルゴリズムを用いた方法では、良品範囲を更新するための演算処理が複雑であるため、各検査対象物の欠陥検査にかかる時間が長くなるという問題がある。   However, a method using a learning type algorithm such as a neural network has a problem in that the time required for defect inspection of each inspection object is increased because the arithmetic processing for updating the non-defective product range is complicated.

本発明は上記事由に鑑みて為されたものであって、比較的簡単な処理で、濃淡値の良品範囲を随時更新して最適化することができる欠陥検査方法および欠陥検査システムを提供することを目的とする。   The present invention has been made in view of the above-described reasons, and provides a defect inspection method and a defect inspection system capable of updating and optimizing a non-defective range of gray values at any time with a relatively simple process. With the goal.

請求項1の発明は、光またはX線を検査対象物に照射した状態で検査対象物の画像を撮像装置にて撮像し、濃淡値を画素値とし且つ同一仕様の多数の検査対象物について同一部位が同一画素に対応する撮像画像を得る撮像過程を含み、前記多数の検査対象物を順次撮像して得られる撮像画像を用いて画像処理装置にて各検査対象物の欠陥の有無を検査する欠陥検査方法であって、検査対象物の撮像画像を設定用画像として設定用画像メモリに記憶する設定用記憶過程と、設定用画像メモリに蓄積された複数枚の設定用画像について画素ごとに濃淡値の度数分布を求める分布算出過程と、度数分布の結果を用い既定の設定ルールに従って濃淡値の良品範囲を設定する範囲設定過程と、撮像画像を撮像画像メモリに記憶する撮像記憶過程と、撮像画像の各画素ごとに濃淡値が前記良品範囲内にあるか否かを判断し、良品範囲内にある画素を良品画素、良品範囲内にない画素を不良画素と判定する画素判定過程と、撮像画像に含まれる不良画素の個数から検査対象物の欠陥の有無を判定する欠陥判定過程と、欠陥判定過程で欠陥がないと判定された場合に当該検査対象物の撮像画像を設定用画像として設定用記憶過程により設定用画像メモリに記憶し、分布算出過程および範囲設定過程により良品範囲を自動的に更新させる範囲更新過程とを有することを特徴とする。   According to the first aspect of the present invention, the image of the inspection object is picked up by the image pickup device in a state where the inspection object is irradiated with light or X-rays. Including an imaging process in which a part has a captured image corresponding to the same pixel, and using the captured images obtained by sequentially capturing the multiple inspection objects, the image processing apparatus inspects each of the inspection objects for defects. A defect inspection method, which is a setting storage process in which a captured image of an inspection target is stored in a setting image memory as a setting image, and a plurality of setting images accumulated in the setting image memory are shaded for each pixel. A distribution calculation process for obtaining a frequency distribution of values, a range setting process for setting a non-defective range of gray values according to a predetermined setting rule using a result of the frequency distribution, an imaging storage process for storing a captured image in a captured image memory, and imaging A pixel determination process for determining whether or not the gray value is within the above-mentioned non-defective range for each pixel of the image, and determining a pixel within the non-defective range as a non-defective pixel and a pixel not within the non-defective range as a defective pixel, and imaging A defect determination process for determining the presence / absence of a defect in the inspection object from the number of defective pixels included in the image, and when the defect determination process determines that there is no defect, a captured image of the inspection object is set as a setting image And a range update process in which the non-defective product range is automatically updated by a distribution calculation process and a range setting process.

この発明によれば、欠陥判定過程で欠陥がないと判定された場合に当該検査対象物の撮像画像を設定用画像として設定用記憶過程により設定用画像メモリに記憶し、分布算出過程および範囲設定過程により良品範囲を自動的に更新させる範囲更新過程を有するので、濃淡値の良品範囲は随時更新されることになる。すなわち、濃淡値の良品範囲は、欠陥判定過程での判定結果がフィードバックされることにより自動的に更新されるので、多数の検査対象物を順次検査する場合でも、常に最適な良品範囲を用いて欠陥の検査を行うことができる。しかも、良品範囲の更新は、設定用画像メモリに蓄積された複数枚の設定用画像について画素ごとに濃淡値の度数分布を求める分布算出過程と、度数分布の結果を用い既定の設定ルールに従って濃淡値の良品範囲を設定する範囲設定過程とで実現されるので、ニューラルネットワーク等の学習型アルゴリズムを用いた方法に比較して、良品範囲を更新するための処理が簡単になる。結果的に、比較的簡単な処理で、濃淡値の良品範囲を随時更新して最適化することができるという利点がある。   According to the present invention, when it is determined that there is no defect in the defect determination process, the captured image of the inspection object is stored as a setting image in the setting image memory by the setting storage process, and the distribution calculation process and the range setting are performed. Since the non-defective range is automatically updated by the process, the non-defective range of the gray value is updated as needed. In other words, the non-defective range of the gray value is automatically updated by feeding back the determination result in the defect determination process, so even when sequentially inspecting a large number of inspection objects, always use the optimal non-defective range. Defect inspection can be performed. In addition, the non-defective product range is updated in accordance with a distribution calculation process for obtaining a frequency distribution of gray values for each pixel of a plurality of setting images stored in the setting image memory, and in accordance with a predetermined setting rule using the result of the frequency distribution. Since it is realized by a range setting process for setting a non-defective range of values, processing for updating the non-defective range is simplified as compared with a method using a learning type algorithm such as a neural network. As a result, there is an advantage that the non-defective range of the gray value can be updated and optimized at any time with a relatively simple process.

請求項2の発明は、請求項1の発明において、前記画素判定過程で前記良品画素と判定された画素のうち前記不良画素と判定された画素で周囲を囲まれたものがある場合、良品画素と判定された当該画素については判定結果を不良画素と改める判定修正過程を有することを特徴とする。   According to a second aspect of the present invention, in the first aspect of the invention, when there is a pixel surrounded by the pixel determined as the defective pixel among the pixels determined as the good pixel in the pixel determination process, The pixel determined to have a determination correction process for changing the determination result to a defective pixel.

この発明によれば、一旦は良品画素と判定された画素であっても周囲が不良画素で囲まれたものについては不良画素と判定結果が改められるので、撮像画像の欠陥部分に含まれる1画素が誤って良品画素と判定された場合に当該画素を不良画素と判定し直すことができる。結果的に、良品画素と不良画素との判定確度が向上するという利点がある。   According to the present invention, even if a pixel is determined to be a non-defective pixel, the determination result is corrected for a pixel surrounded by a defective pixel, so that one pixel included in the defective portion of the captured image is corrected. Is erroneously determined to be a non-defective pixel, the pixel can be determined as a defective pixel. As a result, there is an advantage that the determination accuracy between a non-defective pixel and a defective pixel is improved.

請求項3の発明は、請求項2の発明において、前記判定修正過程において判定結果が前記不良画素と改められた画素がある場合に、前記良品範囲について範囲を狭めるように修正する範囲修正過程を有することを特徴とする。   According to a third aspect of the present invention, in the second aspect of the invention, when there is a pixel whose determination result is changed to the defective pixel in the determination correction process, a range correction process for correcting the non-defective range so as to narrow the range is performed. It is characterized by having.

この発明によれば、撮像画像の欠陥部分に含まれる1画素が誤って良品画素と判定された場合に良品範囲が狭められるので、以降の画素判定過程では、良品画素と認めるための判断基準が厳しくなり、不良画素が誤って良品画素と判定されることを低減できる。   According to the present invention, when one pixel included in the defective portion of the captured image is erroneously determined to be a non-defective pixel, the non-defective range is narrowed. Therefore, in the subsequent pixel determination process, a determination criterion for recognizing a non-defective pixel is It becomes severe, and it can reduce that a defective pixel is erroneously determined as a good pixel.

請求項4の発明は、請求項2の発明において、前記判定修正過程において判定結果が前記不良画素と改められた前記撮像画像が2枚以上の規定枚数に達した場合に、前記良品範囲について範囲を狭めるように修正する範囲修正過程を有することを特徴とする。   According to a fourth aspect of the present invention, in the second aspect of the present invention, the range of the non-defective product range is determined when the number of the captured images in which the determination result is changed to the defective pixel in the determination correction process reaches a specified number of two or more. It has a range correction process for correcting so as to narrow the range.

この発明によれば、撮像画像の欠陥部分に含まれる1画素が誤って良品画素と判定された場合に良品範囲が狭められるので、以降の画素判定過程では、良品画素と認めるための判断基準が厳しくなり、不良画素が誤って良品画素と判定されることを低減できる。しかも、判定結果が改められた撮像画像が規定枚数に達するまでは良品範囲が修正されることはないため、良品画素と判定されるべき画素について判定結果が不良画素に誤って改められたような場合に、良品範囲まで修正されてしまうことを回避できる。すなわち、良品範囲は反復性のあるデータに基づいて修正されるため、良品範囲の信頼性が高くなる。   According to the present invention, when one pixel included in the defective portion of the captured image is erroneously determined to be a non-defective pixel, the non-defective range is narrowed. Therefore, in the subsequent pixel determination process, a determination criterion for recognizing a non-defective pixel is It becomes severe, and it can reduce that a defective pixel is erroneously determined as a good pixel. In addition, the non-defective range is not corrected until the number of captured images with the revised determination result reaches the specified number, so that the determination result for a pixel that should be determined to be a non-defective pixel is erroneously changed to a defective pixel. In this case, it is possible to avoid the correction to the non-defective range. That is, since the good product range is corrected based on repetitive data, the reliability of the good product range becomes high.

請求項5の発明は、請求項1ないし請求項4のいずれかの発明において、前記設定用記憶過程では、前記設定用画像メモリに蓄積された前記設定用画像が設定用画像メモリに記憶可能な上限枚数に達すると、設定用画像メモリ内の設定用画像を古い側から順に削除しながら新しい設定用画像を記憶することを特徴とする。   According to a fifth aspect of the present invention, in the invention according to any one of the first to fourth aspects, the setting image stored in the setting image memory can be stored in the setting image memory in the setting storage process. When the upper limit number is reached, new setting images are stored while the setting images in the setting image memory are deleted in order from the oldest side.

この発明によれば、設定用画像メモリに上限枚数まで設定用画像が記憶されると、以降は古い設定用画像を削除しながら新しい設定用画像を記憶するので、設定用画像メモリには常に新しい設定用画像が残ることとなり、これらの新しい設定用画像を利用して濃淡値の良品範囲を設定することができる。したがって、設定用画像メモリの空き容量がなくなった後も、良品範囲を随時更新して最適化することができる。   According to the present invention, when the setting image is stored in the setting image memory up to the upper limit number, the new setting image is stored while deleting the old setting image, so that the setting image memory is always new. The setting image remains, and the non-defective range of the gray value can be set using these new setting images. Therefore, the non-defective product range can be updated and optimized as needed even after the setting image memory has no free space.

請求項6の発明は、光またはX線を検査対象物に照射する照射装置と、検査対象物の画像を撮像して、濃淡値を画素値とし且つ同一仕様の多数の検査対象物について同一部位が同一画素に対応する撮像画像を得る撮像装置と、前記多数の検査対象物を順次撮像して得られる撮像画像を用いて各検査対象物の欠陥の有無を検査する画像処理装置とを備えた欠陥検査システムであって、画像処理装置が、検査対象物の撮像画像を設定用画像として記憶する設定用画像メモリと、設定用画像メモリに蓄積された複数枚の設定用画像について画素ごとに濃淡値の度数分布を求める分布算出手段と、度数分布の結果を用い既定の設定ルールに従って濃淡値の良品範囲を設定する範囲設定手段と、撮像画像を記憶する撮像画像メモリと、撮像画像の各画素ごとに濃淡値が前記良品範囲内にあるか否かを判断し、良品範囲内にある画素を良品画素、それ以外の画素を不良画素と判定する画素判定手段と、撮像画像に含まれる不良画素の個数から検査対象物の欠陥の有無を判定する欠陥判定手段と、欠陥判定手段で欠陥がないと判定された場合に当該検査対象物の撮像画像を設定用画像として設定用画像メモリに記憶し、分布算出手段および範囲設定手段により良品範囲を自動的に更新させる範囲更新手段とを有することを特徴とする。   The invention according to claim 6 is an irradiation device that irradiates an inspection target with light or X-rays, an image of the inspection target, the gray value is a pixel value, and the same part for a large number of inspection targets having the same specifications Provided with an image pickup device that obtains a picked-up image corresponding to the same pixel, and an image processing device that inspects for the presence or absence of a defect in each inspection target object using captured images obtained by sequentially imaging the numerous inspection target objects. A defect inspection system, in which an image processing apparatus stores a captured image of an inspection object as a setting image, and a plurality of setting images accumulated in the setting image memory for each pixel. A distribution calculating means for obtaining a frequency distribution of values; a range setting means for setting a non-defective range of gray values according to a predetermined setting rule using a result of the frequency distribution; a captured image memory for storing a captured image; and each pixel of the captured image And a pixel determining means for determining whether or not the gray value is within the non-defective range, determining a pixel within the non-defective range as a non-defective pixel, and determining other pixels as defective pixels, and defective pixels included in the captured image A defect determination means for determining the presence or absence of a defect in the inspection object from the number of the images, and when the defect determination means determines that there is no defect, the captured image of the inspection object is stored in the setting image memory as a setting image. And a range updating means for automatically updating the non-defective product range by the distribution calculating means and the range setting means.

この構成によれば、範囲更新手段は、欠陥判定手段での判定結果をフィードバックして、分布算出手段および良品範囲設定手段により濃淡値の良品範囲を自動的に更新させるので、比較的簡単な処理で、濃淡値の良品範囲を随時更新して最適化することができるという利点がある。   According to this configuration, the range update unit feeds back the determination result of the defect determination unit and automatically updates the non-defective range of the gray value by the distribution calculation unit and the non-defective range setting unit. Therefore, there is an advantage that the non-defective product range of the gray value can be updated and optimized at any time.

本発明は、欠陥の有無の判定結果をフィードバックして、複数枚の設定用画像について画素ごとに濃淡値の度数分布を求めその結果を用い既定の設定ルールに従って濃淡値の良品範囲を更新するので、ニューラルネットワーク等の学習型アルゴリズムを用いる場合よりも簡単な処理で、濃淡値の良品範囲を随時更新して最適化することができるという利点がある。   The present invention feeds back the determination result of the presence / absence of a defect, obtains the frequency distribution of the gray value for each pixel of the plurality of setting images, and updates the good / non-defective range of the gray value according to a predetermined setting rule using the result. There is an advantage that the non-defective product range of the gray value can be updated and optimized at any time with simpler processing than when a learning type algorithm such as a neural network is used.

本発明の実施形態1の概略システム構成図である。1 is a schematic system configuration diagram of Embodiment 1 of the present invention. 同上の機能を説明する概略図である。It is the schematic explaining the function same as the above. 同上で用いる度数分布図である。It is a frequency distribution diagram used in the same as the above. 同上の機能を示し、(a)は撮像画像の概略図、(b)は2値化画像の概略図である。The functions of the above are shown, (a) is a schematic diagram of a captured image, and (b) is a schematic diagram of a binarized image. 同上の機能を説明する概略図である。It is the schematic explaining the function same as the above. 同上の動作を示すフローチャートである。It is a flowchart which shows operation | movement same as the above. 同上の範囲更新過程のフローチャートである。It is a flowchart of the range update process same as the above. 同上の説明に用いる撮像画像の概略図である。It is the schematic of the captured image used for description same as the above. 本発明の実施形態2の機能を説明する概略図である。It is the schematic explaining the function of Embodiment 2 of this invention. 同上の機能を説明する概略図である。It is the schematic explaining the function same as the above. 同上の概略システム構成図である。It is a schematic system block diagram same as the above. 同上の動作を示すフローチャートである。It is a flowchart which shows operation | movement same as the above. 同上の範囲修正過程のフローチャートである。It is a flowchart of the range correction process same as the above.

(実施形態1)
本実施形態の欠陥検査システムは、同一仕様の多数の電子部品材料を検査対象とし、各検査対象物の内部の欠陥(異物混入、気泡等)の有無を検査するものである。ここでは、一例として、IC(集積回路)のパッケージ材料となるペレットを検査対象とするが、これに限るものではなく、その他の電子部品材料、あるいは電子部品材料以外の各種物品を検査対象としてもよい。
(Embodiment 1)
The defect inspection system of the present embodiment uses a large number of electronic component materials having the same specifications as inspection targets, and inspects the presence or absence of defects (foreign matter contamination, bubbles, etc.) inside each inspection target object. Here, as an example, pellets that are package materials for ICs (integrated circuits) are to be inspected. However, the present invention is not limited to this, and other electronic component materials or various articles other than electronic component materials may be inspected. Good.

欠陥検査システムは、図1に示すように、検査対象物4の上方に設置され上面側から検査対象物4にX線を照射するX線源からなる照射装置2と、検査対象物4の下方に設置され検査対象物4の画像を順次撮像するX線エリアセンサ(イメージセンサ)からなる撮像装置3と、撮像装置3で得られた撮像画像を用いて検査対象物4の欠陥検査を行う画像処理装置1と、画像処理装置1での検査結果を表示する表示装置5とを備えている。ここでいう撮像画像は、照射装置2からのX線が検査対象物4を透過することによって撮像されるX線透過画像であって、X線の透過強度が高い部位(白く映る部位)ほど高くなる濃淡値を画素値とする濃淡画像である。   As shown in FIG. 1, the defect inspection system is provided above the inspection object 4, an irradiation apparatus 2 including an X-ray source that irradiates the inspection object 4 from the upper surface side, and a lower part of the inspection object 4. An image pickup device 3 composed of an X-ray area sensor (image sensor) that is sequentially installed to pick up an image of the inspection object 4 and an image for performing defect inspection of the inspection object 4 using the picked-up image obtained by the image pickup device 3 A processing device 1 and a display device 5 for displaying inspection results in the image processing device 1 are provided. The captured image here is an X-ray transmission image that is captured when X-rays from the irradiation device 2 pass through the inspection object 4, and the higher the X-ray transmission intensity (the part that appears white) is higher. This is a grayscale image having a grayscale value as a pixel value.

ここに、画像処理装置1で用いられる撮像画像は、同一仕様の検査対象物4については同一部位が同一画素に対応するように位置合わせが為されている。すなわち、同一仕様の多数の検査対象物4を順次撮像した場合、撮像画像内における検査対象物4が占める領域は全ての撮像画像において同一となる。上記位置合わせは、たとえば撮像装置3に対する検査対象物4の位置および向きを全ての検査対象物4について精確に揃えることにより実現できる。また、検査対象物4に基準となるアライメントマークを付加しておけば、撮像装置3で撮像された画像から、前記アライメントマークが同一画素となるように検査対象物4を含む領域を切り出し、切り出した領域を撮像画像として用いることも可能である。   Here, the captured image used in the image processing apparatus 1 is aligned so that the same part corresponds to the same pixel for the inspection object 4 having the same specification. That is, when a large number of inspection objects 4 having the same specifications are sequentially imaged, the area occupied by the inspection object 4 in the captured image is the same in all captured images. The alignment can be realized by, for example, accurately aligning the position and orientation of the inspection object 4 with respect to the imaging device 3 for all the inspection objects 4. If an alignment mark serving as a reference is added to the inspection object 4, a region including the inspection object 4 is cut out from the image captured by the imaging device 3 so that the alignment mark is the same pixel. It is also possible to use the area as a captured image.

照射装置2にはX線をオンオフ制御する照射制御装置6が接続され、撮像装置3には撮像タイミングを制御する撮像制御装置7が接続されており、照射制御装置6と撮像制御装置7とは互いに接続されてX線の照射タイミングと撮像タイミングとの同期をとっている。   An irradiation control device 6 that controls on / off of X-rays is connected to the irradiation device 2, and an imaging control device 7 that controls imaging timing is connected to the imaging device 3. The irradiation control device 6 and the imaging control device 7 are Connected to each other, the X-ray irradiation timing and imaging timing are synchronized.

画像処理装置1は、撮像装置3で得られた撮像画像のデータを保存する撮像画像メモリ11と、撮像画像の各画素ごとに濃淡値が後述する良品範囲内にあるか否かを判定する画素判定手段12と、画素判定手段12の判定結果に基づいて検査対象物4の欠陥の有無を判定する欠陥判定手段13とを具備している。   The image processing apparatus 1 includes a captured image memory 11 that stores data of a captured image obtained by the imaging apparatus 3, and pixels that determine whether or not the gray value is within a non-defective range described later for each pixel of the captured image. A determination unit 12 and a defect determination unit 13 that determines the presence or absence of a defect in the inspection object 4 based on the determination result of the pixel determination unit 12 are provided.

画素判定手段12は、濃淡値が良品範囲内にある画素については良品画素、良品範囲内にない画素については不良画素と判定する。欠陥判定手段13は、撮像画像に含まれる不良画素の個数を所定の閾値(以下、「不良数閾値」という)と比較し、不良画素の個数が不良数閾値以上であった検査対象物4については不良品(欠陥有り)、不良画素の個数が不良数閾値未満であれば良品(欠陥なし)と判定する。欠陥判断手段13での判定結果は、表示装置5に表示されることによって検査員等に報知される。   The pixel determination means 12 determines that a pixel whose gray value is within the non-defective range is a non-defective pixel, and a pixel that is not within the non-defective range is a defective pixel. The defect determination unit 13 compares the number of defective pixels included in the captured image with a predetermined threshold (hereinafter referred to as “defect number threshold”), and the inspection object 4 whose number of defective pixels is equal to or greater than the defect number threshold. Is determined to be a non-defective product (having a defect), and a non-defective product (having no defect) if the number of defective pixels is less than the defect count threshold. The result of determination by the defect determination means 13 is displayed on the display device 5 so as to be notified to an inspector or the like.

ここで、画素判定手段12での判定に用いられる良品範囲は、以下のように設定される。   Here, the non-defective range used for the determination by the pixel determination unit 12 is set as follows.

すなわち、画像処理装置1は、欠陥のない検査対象物4の撮像画像を設定用画像として当該設定用画像のデータを複数保存する設定用画像メモリ14と、設定用画像メモリ14に蓄積された複数枚の設定用画像を用いて良品範囲を設定する演算処理器15とを有している。設定用画像メモリ14は、記憶可能な設定用画像の枚数が最大N枚(Nは整数)に制限されている。本実施形態では、検査対象物4の欠陥検査を開始する前に、図2に示すように欠陥のない検査対象物4を撮像することにより得られたN枚の設定用画像Ic,Ic,・・・,Icが予め設定用画像メモリ14に保存されているものとする。なお、図2においては、ハッチングの種類によって画素ごとの濃淡値を表すものとする(図4,5においても同様)。 That is, the image processing apparatus 1 includes a setting image memory 14 that stores a plurality of setting image data using a captured image of the inspection object 4 having no defect as a setting image, and a plurality of images stored in the setting image memory 14. And an arithmetic processing unit 15 that sets a non-defective range using a set image. The setting image memory 14 is limited to a maximum of N setting images that can be stored (N is an integer). In this embodiment, before starting the defect inspection of the inspection object 4, N setting images Ic 1 and Ic 2 obtained by imaging the inspection object 4 having no defect as shown in FIG. ,..., Ic N are stored in the setting image memory 14 in advance. In FIG. 2, the shade value for each pixel is represented by the type of hatching (the same applies to FIGS. 4 and 5).

演算処理器15は、設定用画像メモリ14に蓄積されたN枚の設定用画像について画素ごとに濃淡値の度数分布を求める分布算出手段15a、並びに、分布算出手段15aにより求めた度数分布データを用い、既定の設定ルールに従って濃淡値の良品範囲を設定する範囲設定手段15bとしての機能を有している。   The arithmetic processor 15 obtains the frequency distribution of the gray value for each pixel for the N setting images stored in the setting image memory 14, and the frequency distribution data obtained by the distribution calculating means 15a. It has a function as range setting means 15b for setting a non-defective range of gray value according to a predetermined setting rule.

要するに、分布算出手段15aでは、設定用画像メモリ14内の設定用画像の各画素ごとに図3に示すように濃淡値の度数分布を求め、同一画素におけるデータ(濃淡値)のばらつきを定量化する。図3は、図2に示す設定用画像Ic,Ic,・・・,Icのある画素P(3,3)についての濃淡値の度数分布を示すヒストグラム(度数分布図)であって、横軸に濃淡値、縦軸に度数を表している。 In short, the distribution calculation means 15a obtains the frequency distribution of the gray value for each pixel of the setting image in the setting image memory 14 as shown in FIG. 3, and quantifies the variation of the data (gray value) in the same pixel. To do. FIG. 3 is a histogram (frequency distribution diagram) showing the frequency distribution of gray values for the pixel P (3, 3) having the setting images Ic 1 , Ic 2 ,..., Ic N shown in FIG. The horizontal axis represents the gray value, and the vertical axis represents the frequency.

範囲設定手段15bは、各画素ごとの度数分布結果から、当該画素における濃淡値の良品範囲を設定ルールに従って自動的に設定する。ここで、良品範囲は各画素において濃淡値のばらつきが良品画素として許容される範囲を規定するものであって、設定ルールの具体例は後述するが、大よそ度数分布図においてピーク(最頻値)となる濃淡値を含むその前後の濃淡値がばらついている範囲が良品範囲に設定される。しかして、図2に示す設定用画像の画素P(3,3)に関しては、図3のように濃淡値のばらつきが生じている範囲が良品範囲として設定される。   The range setting unit 15b automatically sets the non-defective range of the gray value in the pixel from the frequency distribution result for each pixel according to the setting rule. Here, the non-defective range defines a range in which variation in gray value is allowed as a non-defective pixel in each pixel, and a specific example of the setting rule will be described later. The range in which the gray value before and after that including the gray value of) varies is set as the non-defective range. Therefore, with respect to the pixel P (3, 3) of the setting image shown in FIG. 2, a range in which the gray value varies as shown in FIG. 3 is set as a non-defective range.

演算処理器15では、上述した処理を全ての画素についてそれぞれ行うことにより、各画素ごとに濃淡値の良品範囲を個別に設定する。本実施形態では、設定された良品範囲は、演算処理器15内のメモリ(図示せず)に画素ごとに記憶される。各画素についての良品範囲は、当該範囲の下限となる濃淡値(以下、「良品下限値」という)および当該範囲の上限となる濃淡値(以下、「良品上限値」という)によって定められる。   The arithmetic processing unit 15 individually sets the non-defective range of the gray value for each pixel by performing the above-described processing for all the pixels. In the present embodiment, the set non-defective range is stored for each pixel in a memory (not shown) in the arithmetic processor 15. The non-defective range for each pixel is determined by a gray value (hereinafter referred to as “non-defective lower limit value”) that is the lower limit of the range and a gray value (hereinafter referred to as “non-defective upper limit value”) that is the upper limit of the range.

このようにして設定された良品範囲を用いることにより、画素判定手段12では、良品画素と不良画素との判定が行われ、良品画素を画素値「0」、不良画素を画素値「1」とした2値化画像が生成される。たとえば、図4(a)に示す撮像画像Imにおける画素P(3,3)の濃淡値E(3,3)が図3の良品範囲内(良品下限値RD−良品上限値RU間)になければ、当該画素P(3,3)が不良画素と判定され、図4(b)のように当該画素P(3,3)の画素値を「1」(図中黒塗り部分)とする2値化画像が生成される。   By using the non-defective range set in this way, the pixel determination unit 12 determines whether the non-defective pixel is defective or defective, and sets the non-defective pixel to a pixel value “0” and the defective pixel to a pixel value “1”. A binarized image is generated. For example, the gray value E (3, 3) of the pixel P (3, 3) in the captured image Im shown in FIG. 4A must be within the non-defective range shown in FIG. 3 (between the good product lower limit RD and the good product upper limit RU). For example, the pixel P (3, 3) is determined to be a defective pixel, and the pixel value of the pixel P (3, 3) is set to “1” (black portion in the figure) as shown in FIG. A valued image is generated.

ところで、上記構成の欠陥検査システムでは、同一仕様の多数の検査対象物4を順次検査するに当たって、たとえば照射装置2の出力低下などによって撮像画像の濃淡値が変化することがあるので、最初に予め設定される良品範囲が必ずしも最適な値とは限らず、良品範囲を随時更新して最適化することが望ましい。そこで、本実施形態では画像処理装置1に以下の構成を採用することにより、濃淡値の良品範囲が随時更新されるようにしてある。   By the way, in the defect inspection system having the above-described configuration, when sequentially inspecting a large number of inspection objects 4 having the same specification, the gray value of the captured image may change due to, for example, a decrease in output of the irradiation device 2. The set good product range is not necessarily an optimum value, and it is desirable to update the good product range as needed to optimize it. Thus, in the present embodiment, the following configuration is adopted in the image processing apparatus 1 so that the non-defective range of the gray value is updated as needed.

すなわち、画像処理装置1は、欠陥判定手段13で良品(欠陥なし)と判定された場合に、当該検査対象物4の撮像画像を設定用画像として設定用画像メモリ14に追加する範囲更新手段10を有している。ここで、演算処理器15は、予め設定用画像メモリ14に記憶されている設定用画像から良品範囲を設定するだけでなく、範囲更新手段10により新たな設定用画像が設定用画像メモリ14に保存される度に、各画素ごとの濃淡値の度数分布を再び求め、当該度数分布結果から良品範囲を再設定するものとする。しかして、欠陥判定手段13で良品と判定される度、範囲更新手段10は、設定用画像を設定用画像メモリ14に追加し、追加された設定用画像を含む複数枚の設定用画像に基づいて演算処理器15に良品範囲を自動的に更新させることとなる。   That is, the image processing apparatus 1 adds the captured image of the inspection object 4 to the setting image memory 14 as a setting image when the defect determining unit 13 determines that the product is non-defective (no defect). have. Here, the arithmetic processor 15 not only sets a non-defective range from the setting image stored in the setting image memory 14 in advance, but also sets a new setting image in the setting image memory 14 by the range update means 10. Each time the image is stored, the frequency distribution of the gray value for each pixel is obtained again, and the non-defective range is reset from the frequency distribution result. Thus, each time the defect determination unit 13 determines that the product is non-defective, the range update unit 10 adds the setting image to the setting image memory 14 and based on the plurality of setting images including the added setting image. Thus, the non-defective product range is automatically updated by the processor 15.

ここにおいて、設定用画像メモリ14に蓄積された設定用画像が設定用画像メモリ14に記憶可能な上限枚数に達すると、以降は、設定用画像メモリ14内の設定用画像を古い側から順に削除しながら新しい設定用画像を設定用画像メモリ14に記憶する。つまり、図5(a)のように設定用画像メモリ14にN枚の設定用画像Ic,Ic,・・・,Icが既に記憶されている状態で、欠陥判定手段13で良品と判定された検査対象物4の撮像画像Imを新たに設定用画像として設定用画像メモリ14に記憶する場合、図5(b)のように、設定用画像メモリ14内の最も古い設定用画像Icは削除され、撮像画像Imが記憶される。したがって、設定用画像メモリ14内には、常に新しい方から順にN枚の設定用画像が残ることとなり、演算処理器15では、これら最新のN枚の設定用画像に基づいて良品範囲を更新することができる。 Here, when the number of setting images stored in the setting image memory 14 reaches the upper limit number that can be stored in the setting image memory 14, the setting images in the setting image memory 14 are deleted in order from the oldest side. The new setting image is stored in the setting image memory 14. That is, as shown in FIG. 5 (a), in the state in which N setting images Ic 1 , Ic 2 ,..., Ic N have already been stored in the setting image memory 14, When the determined captured image Im of the inspection object 4 is newly stored in the setting image memory 14 as a setting image, as shown in FIG. 5B, the oldest setting image Ic in the setting image memory 14 is stored. N is deleted and the captured image Im is stored. Accordingly, N setting images always remain in order from the newest in the setting image memory 14, and the arithmetic processor 15 updates the non-defective range based on these latest N setting images. be able to.

以下、上記構成の欠陥検査システムを用いた検査対象物4の欠陥検査方法について図6および図7に示すフローチャートを参照して説明する。ここでは、図8に示すようにp画素×r画素(=横×縦)の濃淡画像を撮像画像Imとし、各画素は、画像の横方向の座標位置i(=1,2,・・・,p)と縦方向の座標位置j(=1,2,・・・,r)とを用いてP(i,j)として表されるものとする。なお、フローチャート中のE(i,j)は撮像画像Imにおける画素P(i,j)の濃淡値を表し、RD(i,j)は画素P(i,j)の良品下限値、RU(i,j)は画素P(i,j)の良品上限値をそれぞれ表し、CH(i,j)は画素P(i,j)についての良品画素・不良画素の判定結果(良品画素は「0」、不良画素は「1」とする)を表している。   Hereinafter, a defect inspection method for the inspection object 4 using the defect inspection system having the above configuration will be described with reference to flowcharts shown in FIGS. 6 and 7. Here, as shown in FIG. 8, a grayscale image of p pixels × r pixels (= horizontal × vertical) is taken as a captured image Im, and each pixel has a horizontal coordinate position i (= 1, 2,... , P) and the coordinate position j (= 1, 2,..., R) in the vertical direction, and expressed as P (i, j). In the flowchart, E (i, j) represents the gray value of the pixel P (i, j) in the captured image Im, RD (i, j) represents the good product lower limit value of the pixel P (i, j), RU ( i, j) represents the non-defective upper limit value of the pixel P (i, j), and CH (i, j) is a non-defective pixel / defective pixel determination result for the pixel P (i, j) (non-defective pixel is “0”). ", The defective pixel is" 1 ").

検査が開始すると、画像処理装置1は、まず撮像装置3で検査対象物4を撮像する(撮像過程)ことにより得られた撮像画像Imを撮像画像メモリ11に記憶する(撮像記憶過程)。それから、画素判定手段12では撮像画像Imの各画素P(i,j)ごとに濃淡値E(i,j)が良品範囲内にあるか否かを判断し、良品範囲内にあれば良品画素、良品範囲内になければ不良画素と判定する(画素判定過程S3〜S5)。ここで、画素判定過程は撮像画像Imの全画素について繰り返される(S1,S2,S6,S7)。   When the inspection starts, the image processing apparatus 1 first stores the captured image Im obtained by imaging the inspection object 4 with the imaging device 3 (imaging process) in the captured image memory 11 (imaging storage process). Then, the pixel determination unit 12 determines whether or not the gray value E (i, j) is within the non-defective range for each pixel P (i, j) of the captured image Im. If it is not within the non-defective range, it is determined as a defective pixel (pixel determination steps S3 to S5). Here, the pixel determination process is repeated for all the pixels of the captured image Im (S1, S2, S6, S7).

その後、欠陥判定手段13では、各撮像画像Imに含まれる不良画素の個数を不良画素数DTとしてそれぞれカウントし(S8〜S14)、当該不良画素数DTを不良数閾値と比較することにより、検査対象物4の良品・不良品の判定を行う(欠陥判定過程S15〜S17)。このとき、検査対象物4が良品と判定されると、濃淡値の良品範囲を更新するための処理(範囲更新過程S18)に移行する。   Thereafter, the defect determination unit 13 counts the number of defective pixels included in each captured image Im as the number of defective pixels DT (S8 to S14), and compares the number of defective pixels DT with a defect number threshold value, thereby performing inspection. Non-defective / defective products of the object 4 are determined (defect determination processes S15 to S17). At this time, if the inspection object 4 is determined to be a non-defective product, the process proceeds to a process (range update process S18) for updating the non-defective product range of the gray value.

範囲更新過程ではまず、図7に示すように範囲更新手段10が良品と判定された検査対象物4の撮像画像Imを設定用画像メモリ14に記憶する(良品記憶過程S22〜S25)。この良品記憶過程では、上述したように設定用画像メモリ14内の最も古い画像が削除され、前記良品の撮像画像Imが設定用画像メモリ14に追加される。なお、図7中のC(k)(i,j)は、設定用画像メモリ14内の新しい側からk(=1,2,・・・,N)番目の設定用画像Icにおける画素P(i,j)の濃淡値を表している。 In the range update process, first, as shown in FIG. 7, the range update means 10 stores the captured image Im of the inspection object 4 determined to be non-defective in the setting image memory 14 (non-defective product storage steps S22 to S25). In this non-defective product storage process, as described above, the oldest image in the setting image memory 14 is deleted, and the non-defective captured image Im is added to the setting image memory 14. Note that C (k) (i, j) in FIG. 7 is a pixel P in the k (= 1, 2,..., N) -th setting image Ic k from the new side in the setting image memory 14. It represents the gray value of (i, j).

そして、分布算出手段15aは、設定用画像メモリ14に蓄積されたN枚の設定用画像について画素P(i,j)の濃淡値の度数分布を求める(分布算出過程S26)。ここに、S26では濃淡値の平均を数1に示す式によって求め、さらに、濃淡値の標準偏差を数2に示す式により求めている。   Then, the distribution calculating unit 15a obtains the frequency distribution of the gray value of the pixel P (i, j) for the N setting images stored in the setting image memory 14 (distribution calculating step S26). Here, in S26, the average of the gray value is obtained by the equation shown in Equation 1, and the standard deviation of the light and shade value is obtained by the equation shown in Equation 2.

Figure 2010230452
Figure 2010230452

Figure 2010230452
Figure 2010230452

範囲設定手段15bは、度数分布結果を用いて既定の設定ルールに従って画素P(i,j)における濃淡値の良品範囲を再設定する(範囲設定過程S27)。ここで、設定ルールは、S26で求めた濃淡値の平均および標準偏差を用い、数3および数4に示す式によって良品下限値RD(i,j)および良品上限値RU(i,j)を決めるものとする。図7の例では、S22〜S27の処理は各画素P(i,j)ごとに為され、これらの処理が撮像画像Imの全画素について繰り返される(S20,S21,S28,S29)。   The range setting unit 15b resets the non-defective range of the gray value in the pixel P (i, j) according to a predetermined setting rule using the frequency distribution result (range setting step S27). Here, the setting rule uses the average and standard deviation of the gray values obtained in S26, and calculates the non-defective product lower limit value RD (i, j) and the non-defective product upper limit value RU (i, j) by the equations shown in Equations 3 and 4. It shall be decided. In the example of FIG. 7, the processes of S22 to S27 are performed for each pixel P (i, j), and these processes are repeated for all the pixels of the captured image Im (S20, S21, S28, S29).

Figure 2010230452
Figure 2010230452

Figure 2010230452
Figure 2010230452

上述した図6のS1〜S18の処理は、撮像過程により新たな検査対象物4の撮像画像Imが取得される度に繰り返される。   The above-described processes of S1 to S18 in FIG. 6 are repeated each time a captured image Im of a new inspection object 4 is acquired by the imaging process.

以上説明した欠陥検査方法によれば、欠陥判定手段13にて良品との判定が為される度に、設定用画像メモリ14に新たな設定用画像が追加され、当該設定用画像を含めて各画素の濃淡値の度数分布が再計算されて、度数分布結果により各画素ごとに濃淡値の良品範囲が再設定される。このように濃淡値の良品範囲は欠陥判定手段13での判定結果がフィードバックされることにより自動更新されるので、多数の検査対象物4を順次検査するに当たって、たとえば照射装置2の出力低下などによって撮像画像の濃淡値が変化することがあっても、濃淡値の良品範囲を随時更新して最適化することができる。その結果、常に最適な良品範囲を用いて欠陥の有無を検査できるため、欠陥検査の信頼性が高くなる。   According to the defect inspection method described above, a new setting image is added to the setting image memory 14 every time the defect determining unit 13 determines that the product is non-defective, and each image including the setting image is included. The frequency distribution of the gray value of the pixel is recalculated, and the non-defective range of the gray value is reset for each pixel based on the frequency distribution result. In this way, the non-defective product range of the gray value is automatically updated by feeding back the determination result of the defect determination means 13, and therefore, when sequentially inspecting a large number of inspection objects 4, for example, due to a decrease in the output of the irradiation device 2 or the like Even if the gray value of the captured image changes, the non-defective product range of the gray value can be updated and optimized at any time. As a result, since the presence or absence of a defect can always be inspected using the optimum non-defective product range, the reliability of the defect inspection is increased.

ここで、良品範囲の更新は、濃淡値の度数分布を求め、その結果を用いて良品範囲を再設定するという比較的簡単な処理で実現されるので、たとえばニューラルネットワーク等の学習型アルゴリズムを用いて良品範囲を更新する場合に比べ、良品範囲の更新のための処理を簡略化することができる。   Here, the update of the non-defective range is realized by a relatively simple process of obtaining the frequency distribution of the gray value and resetting the non-defective range using the result, for example, using a learning type algorithm such as a neural network. Therefore, the process for updating the non-defective range can be simplified as compared with the case of updating the non-defective range.

ところで、本実施形態では、検査対象物4の欠陥検査を開始する前に、欠陥のない検査対象物4の撮像画像を設定用画像として設定用画像メモリ14に予め記憶しているが、この例に限らず、欠陥検査の開始前にランダムに選択された複数の検査対象物4の撮像画像を設定用画像として設定用画像メモリ14に予め記憶し、これらの設定用画像から欠陥検査の開始時点の良品範囲が設定されるようにしてもよい。この場合、設定用画像メモリ14に予め記憶される設定用画像には、良品だけでなく不良品となる検査対象物4の撮像画像も含まれるため、画素ごとの濃淡値のばらつきが大きく、欠陥検査の開始時点の良品範囲は比較的広くなる。その後、欠陥検査によって良品と判定された検査対象物4の撮像画像が設定用画像メモリ14に追加されていくに連れて、画素ごとの濃淡値のばらつきが徐々に小さくなって良品範囲の信頼性が向上する。   By the way, in this embodiment, before starting the defect inspection of the inspection object 4, the captured image of the inspection object 4 having no defect is stored in advance in the setting image memory 14 as a setting image. Not limited to this, captured images of a plurality of inspection objects 4 randomly selected before the start of defect inspection are stored in advance in the setting image memory 14 as setting images, and the defect inspection start time is determined from these setting images. The non-defective product range may be set. In this case, the setting image stored in advance in the setting image memory 14 includes not only a non-defective product but also a captured image of the inspection object 4 that is a defective product. The non-defective product range at the start of the inspection is relatively wide. Thereafter, as the captured image of the inspection object 4 determined as a non-defective product by the defect inspection is added to the setting image memory 14, the variation in the gray value for each pixel gradually decreases, and the reliability of the non-defective range is determined. Will improve.

また、設定用画像メモリ14に設定用画像を予め記憶しておくのではなく、たとえば検査員が手入力した良品範囲を欠陥検査の開始時点における良品範囲として用いてもよい。この場合、検査を開始してから設定用画像メモリ14にN枚の設定用画像が蓄積されるまでは、手入力された良品範囲によって欠陥検査を実施し、設定用画像メモリ14にN枚の設定用画像が蓄積された以降、演算処理器15で自動設定された良品範囲を用いることが望ましい。   Further, instead of storing the setting image in the setting image memory 14 in advance, for example, a non-defective range manually input by an inspector may be used as the non-defective range at the start of defect inspection. In this case, from the start of inspection until the N setting images are accumulated in the setting image memory 14, defect inspection is performed based on the manually input non-defective product range, and N setting images are stored in the setting image memory 14. After the setting image is accumulated, it is desirable to use the non-defective range automatically set by the arithmetic processor 15.

なお、上記実施形態では、撮像画像の各画素ごとに良品範囲を設定し良品画素・不良画素の判定を行うものとして説明したが、ここでいう画素は、撮像装置3としてのX線エリアセンサで得られる生画像の1画素に対応するものに限らない。すなわち、たとえば撮像装置3で得られた生画像について複数画素を1単位として区画化した画像を作成し、当該画像の1区画を1画素として、上記良品範囲の設定および良品画素・不良画素の判定を行うようにしてもよい。   In the above embodiment, the non-defective pixel / defective pixel is determined by setting a non-defective range for each pixel of the captured image. However, the pixel here is an X-ray area sensor as the imaging device 3. It is not limited to one corresponding to one pixel of the obtained raw image. That is, for example, an image obtained by partitioning a plurality of pixels as a unit from a raw image obtained by the imaging device 3 is created, and the above-described non-defective range setting and non-defective pixel / defective pixel determination are performed using one section of the image as one pixel. May be performed.

(実施形態2)
本実施形態の欠陥検査システムを用いた欠陥検査方法は、画素判定手段12にて一旦は良品画素と判定された画素について、その判定結果を不良画素に修正する判定修正手段16(図11参照)を画像処理装置1に有する点が実施形態1と相違する。
(Embodiment 2)
In the defect inspection method using the defect inspection system according to the present embodiment, the determination correction unit 16 (see FIG. 11) corrects the determination result to a defective pixel for a pixel that is once determined to be a good pixel by the pixel determination unit 12. Is different from that of the first embodiment.

判定修正手段16は、画素判定手段12で良品画素と判定された画素のうち、図9(a)に示す2値化画像の画素P(4,3)のように、四方(上下左右)を不良画素(図中黒塗り部分)で囲まれたものがあれば、図9(b)に示すように当該画素の判定結果を不良画素に修正する。すなわち、検査対象物4に欠陥がある場合、画素判定手段12では欠陥の全体に亘って不良画素と判定されるはずであるから、不良画素で囲まれた領域に1画素分だけ良品画素が紛れるようなことは通常起こり得ず、当該良品画素は不良画素が誤って良品画素と判定されたものと推定できる。したがって、判定修正手段16で四方を不良画素に囲まれた良品画素を不良画素と改めることにより良品画素・不良画素の判定確度が向上する。   Of the pixels determined to be non-defective pixels by the pixel determination unit 12, the determination correction unit 16 converts four sides (up, down, left, and right) like a pixel P (4, 3) of the binarized image shown in FIG. If there is a pixel surrounded by a defective pixel (black portion in the figure), the determination result of the pixel is corrected to a defective pixel as shown in FIG. 9B. That is, when the inspection object 4 has a defect, the pixel determination unit 12 should determine the defective pixel over the entire defect, so that the non-defective pixel is lost by one pixel in the area surrounded by the defective pixel. Such a situation cannot usually occur, and it can be estimated that the defective pixel is erroneously determined as a defective pixel. Therefore, the non-defective pixel / defective pixel determination accuracy is improved by changing the non-defective pixel surrounded by the defective pixels on all sides by the determination correcting means 16.

欠陥判定手段13では、判定修正手段16で判定結果が修正された後の不良画素の個数から、検査対象物4の良品(欠陥なし)・不良品(欠陥有り)を判定しており、結果的に、検査対象物4の良品・不良品の判定確度が向上し、欠陥検査の信頼性が向上する。   The defect determination unit 13 determines whether the inspection object 4 is non-defective (no defect) or defective (has a defect) from the number of defective pixels after the determination result is corrected by the determination correction unit 16. In addition, the determination accuracy of the non-defective product / defective product of the inspection object 4 is improved, and the reliability of the defect inspection is improved.

ところで、本実施形態では、判定修正手段16により判定結果が不良画素と改められた画素がある場合、良品範囲を狭めるように修正する範囲修正手段15cとしての機能を演算処理器15に有している。   By the way, in the present embodiment, when there is a pixel whose determination result is changed to a defective pixel by the determination correction unit 16, the arithmetic processor 15 has a function as the range correction unit 15c that corrects the non-defective product range to be narrowed. Yes.

たとえば、撮像画像における画素P(4,3)の濃淡値E(4,3)が図10(a)の良品範囲内にあり、画素判定手段12にて画素P(4,3)は良品画素と一旦判定された後、図9のように判定修正手段16にて当該画素P(4,3)の判定結果が不良画素に修正された場合、範囲修正手段15cは良品範囲を修正によって狭くする(図10(b)参照)。この良品範囲の修正は、判定修正手段16で判定結果が修正された画素だけでなく、全ての画素について為される。ここでは、全ての画素について良品範囲の修正量(良品上限値RUあるいは良品下限値RDのシフト量)は一律とされる。これにより、画素判定手段12で誤って良品画素と判定された画素について判定修正手段16で判定結果が改められた場合、以降、画素判定手段12で良品画素と判定されるための判断基準が厳しくなり、同様の誤判定が画素判定手段12で生じにくくなる。   For example, the grayscale value E (4,3) of the pixel P (4,3) in the captured image is within the non-defective range of FIG. 10 (a), and the pixel determination means 12 determines that the pixel P (4,3) is a nondefective pixel. When the determination result of the pixel P (4, 3) is corrected to a defective pixel by the determination correction unit 16 as shown in FIG. 9, the range correction unit 15c narrows the non-defective range by correction. (See FIG. 10B). The correction of the non-defective product range is performed not only for the pixel whose determination result is corrected by the determination correction means 16 but also for all the pixels. Here, the correction amount of the good product range (the shift amount of the good product upper limit value RU or the good product lower limit value RD) is uniform for all the pixels. Thereby, when the determination result is amended by the determination correction unit 16 for the pixel erroneously determined to be a non-defective pixel by the pixel determination unit 12, the criteria for determining the non-defective pixel by the pixel determination unit 12 are stricter thereafter. Thus, the same erroneous determination is less likely to occur in the pixel determination unit 12.

ここに、良品上限値RUが最適値より高かったために不良画素が良品画素と誤判定された場合には良品上限値RUを引き下げ、良品下限値RDが最適値より低かったために不良画素が良品画素と誤判定された場合には良品下限値RDを引き上げることにより良品判定が修正される。いずれの理由で不良画素が良品画素と誤判定されたかは、判定結果が修正された画素の周囲の不良画素の濃淡値が、良品上限値RUを超えているのかあるいは良品下限値RDを下回っているのかによって判断される。すなわち、図10の例では、画素P(4,3)の周囲の不良画素の濃淡値が良品下限値RDを下回っているにもかかわらず、画素P(4,3)の濃淡値E(4,3)が良品下限値RD(4,3)より高かったために画素P(4,3)について良品画素と誤判定されたので、良品下限値RDを引き上げることにより良品範囲を修正する。   If the defective pixel is erroneously determined to be a non-defective pixel because the non-defective upper limit value RU is higher than the optimum value, the non-defective upper limit value RU is lowered, and the non-defective pixel lower limit value RD is lower than the optimum value. Is erroneously determined, the good product determination is corrected by raising the good product lower limit RD. The reason why the defective pixel is erroneously determined to be a non-defective pixel is determined by whether the gray value of the defective pixel around the pixel whose determination result is corrected exceeds the non-defective upper limit RU or falls below the non-defective lower limit RD. It is judged by whether it is. That is, in the example of FIG. 10, the gray value E (4) of the pixel P (4, 3) is low even though the gray value of the defective pixel around the pixel P (4, 3) is below the non-defective lower limit RD. , 3) is higher than the non-defective product lower limit value RD (4, 3), and thus the pixel P (4, 3) is erroneously determined as a non-defective pixel. Therefore, the non-defective product range is corrected by raising the non-defective product lower limit value RD.

また、良品範囲の修正量(良品上限値RUあるいは良品下限値RDのシフト量)は、予め一定量に決めておいてもよいが、判定結果が修正された画素やその周囲の画素の濃淡値に基づいて、その都度決定するようにしてもよい。図10の例では、判定修正手段16にて判定結果が修正された画素P(4,3)の濃淡値E(4,3)が良品範囲外となるように、当該画素の濃淡値に基づいて良品範囲の修正量を決定している。具体的には、画素P(4,3)の良品下限値RD(4,3)と濃淡値E(4,3)との差分だけ、良品下限値RDを引き上げている。   Further, the correction amount of the non-defective range (the shift amount of the non-defective product upper limit value RU or the non-defective product lower limit value RD) may be determined in advance, but the gray value of the pixel whose determination result is corrected and the surrounding pixels. Based on the above, it may be determined each time. In the example of FIG. 10, the gray level value E (4, 3) of the pixel P (4, 3) whose determination result is corrected by the determination correction unit 16 is based on the gray level value of the pixel so that it is outside the non-defective range. The amount of correction for the non-defective range is determined. Specifically, the non-defective product lower limit value RD is increased by the difference between the non-defective product lower limit value RD (4, 3) and the gray value E (4, 3) of the pixel P (4, 3).

ただし、判定修正手段16にて判定結果が修正される度に良品範囲が修正されるものとすると、良品画素と判定されるべき画素が判定修正手段16にて誤って不良画素と改められた場合に、良品範囲まで誤って修正されてしまうという不都合を生じ得る。そこで、本実施形態では、判定修正手段16において判定結果が改められた撮像画像が2枚以上の規定枚数に達した場合にのみ、範囲修正手段15cにて良品範囲の修正を行うものとする。   However, assuming that the non-defective range is corrected every time the determination result is corrected by the determination correction unit 16, a pixel that should be determined to be a non-defective pixel is erroneously changed to a defective pixel by the determination correction unit 16. In addition, there may be a disadvantage that the product is erroneously corrected to the non-defective range. Therefore, in the present embodiment, the non-defective product range is corrected by the range correcting unit 15c only when the number of captured images whose determination result has been corrected by the determination correcting unit 16 has reached a specified number of two or more.

具体的には、図11に示すように、判定修正手段16にて判定結果が修正された撮像画像を修正用画像として記憶する修正用画像メモリ17を画像処理装置1に設け、範囲修正手段15cは、修正用画像メモリ17に記憶された修正用画像が規定枚数に達する度に、これら規定枚数の修正用画像を用いて良品範囲を修正する。修正用画像メモリ17は、設定用画像メモリ14と同様に、蓄積された修正用画像が記憶可能な上限枚数に達すると、以降は、修正用画像を古い側から順に削除しながら新しい修正用画像を記憶する。ここに、範囲修正手段15cは、上記規定枚数の修正用画像から判定修正手段16で判定結果が改められた画素の濃淡値をそれぞれ抽出し、たとえばこれらの濃淡値と良品下限値RDまたは良品上限値RUとの差分の最頻値をとることによって、良品範囲の修正量(良品上限値RUあるいは良品下限値RDのシフト量)を決定する。   Specifically, as shown in FIG. 11, a correction image memory 17 for storing a captured image whose determination result is corrected by the determination correction unit 16 as a correction image is provided in the image processing apparatus 1, and a range correction unit 15c. When the number of correction images stored in the correction image memory 17 reaches the specified number, the non-defective product range is corrected using the specified number of correction images. Similar to the setting image memory 14, when the correction image memory 17 reaches the upper limit number of stored correction images, the correction image memory 17 subsequently deletes the correction images in order from the old side, and then adds new correction images. Remember. Here, the range correction unit 15c extracts the gray values of the pixels whose determination results have been corrected by the determination correction unit 16 from the specified number of correction images, for example, the gray value and the non-defective product lower limit value RD or the non-defective product upper limit value. By taking the mode of the difference from the value RU, the correction amount of the non-defective range (the shift amount of the non-defective product upper limit value RU or the non-defective product lower limit value RD) is determined.

しかして、たとえば良品下限値よりも高かったために良品画素と誤判定された画素が複数あった場合には、これらの画素の濃淡値に基づいて全画素について良品下限値を引き上げることにより良品範囲が修正される。その結果、良品範囲が狭められ、以降の欠陥検査においては同様の誤判定が画素判定手段12で生じにくくなり、欠陥検査の信頼性が向上する。   Thus, for example, when there are multiple pixels that are erroneously determined to be non-defective pixels because they are higher than the non-defective product lower limit value, the non-defective product range is increased by raising the non-defective product lower limit value for all pixels based on the gray value of these pixels. Will be corrected. As a result, the non-defective product range is narrowed, and in the subsequent defect inspection, the same erroneous determination is less likely to occur in the pixel determination means 12, and the reliability of the defect inspection is improved.

以下、上記構成の欠陥検査システムを用いた検査対象物4の欠陥検査方法について図12および図13に示すフローチャートを参照して説明する。ここで、図12のS30〜S36の処理は、実施形態1で説明した図6のフローチャートのS7とS8との間に挿入されるものであって、実施形態1と重複する部分については説明を省略する。   Hereinafter, a defect inspection method for the inspection object 4 using the defect inspection system having the above configuration will be described with reference to flowcharts shown in FIGS. Here, the processes of S30 to S36 in FIG. 12 are inserted between S7 and S8 in the flowchart of FIG. 6 described in the first embodiment, and the description of the same parts as those in the first embodiment will be described. Omitted.

すなわち、判定修正手段16では、画素判定過程S3〜S5において良品画素と判定された画素を対象に、不良画素にて周囲を囲まれたものがあるか否かを判断し、そのような画素があれば当該画素の判定結果を不良画素に修正する(判定修正過程S32,S33)。ここで、判定結果の修正が為されると、良品範囲について範囲を狭めるように修正する処理(範囲修正過程S34)に移行する。   That is, the determination correction unit 16 determines whether there is a pixel surrounded by defective pixels for pixels determined to be non-defective pixels in the pixel determination processes S3 to S5. If there is, the determination result of the pixel is corrected to a defective pixel (determination correction processes S32 and S33). Here, when the determination result is corrected, the process proceeds to a process of correcting the non-defective range so as to narrow the range (range correction step S34).

範囲修正過程ではまず、図13に示すように判定修正手段16にて判定結果が修正された撮像画像を修正用画像として修正用画像メモリ17に記憶する(修正記憶過程S42〜S45)。この修正記憶過程では、修正用画像メモリ17内の最も古い画像が削除され、前記修正用画像が修正用画像メモリ17に追加される。図13中のNG(k)(i,j)は、修正用画像メモリ17内の新しい側からk(=1,2,・・・,M)番目の修正用画像における画素P(i,j)の濃淡値を表している。S42〜S45の処理は、各画素P(i,j)ごとに為され、これらの処理が全画素について繰り返される(S40,S41,S46,S47)。   In the range correction process, first, as shown in FIG. 13, the captured image whose determination result has been corrected by the determination correction means 16 is stored in the correction image memory 17 as a correction image (correction storage processes S42 to S45). In this correction storage process, the oldest image in the correction image memory 17 is deleted, and the correction image is added to the correction image memory 17. NG (k) (i, j) in FIG. 13 is a pixel P (i, j) in the k (= 1, 2,..., M) th correction image from the new side in the correction image memory 17. ). The processes of S42 to S45 are performed for each pixel P (i, j), and these processes are repeated for all the pixels (S40, S41, S46, S47).

そして、修正用画像メモリ17に記憶された修正用画像が規定枚数に達すると(S48〜S50の処理でカウント値CTが規定値CUに達すると)、範囲修正手段15cにより良品範囲が修正される(S53,S54)。図13中のΔRD、ΔRUはそれぞれ良品下限値RDのシフト量、良品上限値RUのシフト量を表している。ここで、良品範囲の修正処理は、全画素について繰り返される(S51,S52,S55,S56)。   When the number of correction images stored in the correction image memory 17 reaches the specified number (when the count value CT reaches the specified value CU in the processing of S48 to S50), the non-defective product range is corrected by the range correction means 15c. (S53, S54). ΔRD and ΔRU in FIG. 13 represent the shift amount of the non-defective product lower limit value RD and the shift amount of the non-defective product upper limit value RU, respectively. Here, the non-defective range correction process is repeated for all pixels (S51, S52, S55, S56).

上述した図12におけるS32〜S34の処理は各画素P(i,j)ごとに為され、これらの処理が撮像画像Imの最外周部(つまり、P(1,j)、P(i,1)、P(p,j)、P(i,r)の4辺)を除く全画素について繰り返される(S30,S31,S35,S36)。   The above-described processes of S32 to S34 in FIG. 12 are performed for each pixel P (i, j), and these processes are performed on the outermost peripheral portion (that is, P (1, j), P (i, 1) of the captured image Im. ), P (p, j), and P (i, r) (4 sides) are repeated for all pixels (S30, S31, S35, S36).

なお、その他の構成および機能は実施形態1と同様である。   Other configurations and functions are the same as those in the first embodiment.

ところで、上記各実施形態では、X線透過画像を用いて検査対象物4の内部の欠陥を検査する欠陥検査システムに本発明を適用する例を示したが、この例に限るものではなく、たとえば検査対象物4に光を照射した状態で照射装置2側から検査対象物4を撮像し、検査対象物4の外観上の欠陥を検査する欠陥検査システムに本発明を適用してもよい。   By the way, in each said embodiment, although the example which applies this invention to the defect inspection system which test | inspects the defect inside the test target object 4 using an X-ray transmission image was shown, it is not restricted to this example, For example, The present invention may be applied to a defect inspection system in which the inspection object 4 is imaged from the irradiation device 2 side while the inspection object 4 is irradiated with light, and defects on the appearance of the inspection object 4 are inspected.

1 画像処理装置
2 照射装置
3 撮像装置
4 検査対象物
10 範囲更新手段
11 撮像画像メモリ
12 画素判定手段
13 欠陥判定手段
14 設定用画像メモリ
15 演算処理器
15a 分布算出手段
15b 範囲設定手段
15c 範囲修正手段
16 判定修正手段
17 修正用画像メモリ
DESCRIPTION OF SYMBOLS 1 Image processing apparatus 2 Irradiation apparatus 3 Imaging apparatus 4 Inspection object 10 Range update means 11 Captured image memory 12 Pixel determination means 13 Defect determination means 14 Setting image memory 15 Arithmetic processor 15a Distribution calculation means 15b Range setting means 15c Range correction Means 16 Determination / correction means 17 Image memory for correction

Claims (6)

光またはX線を検査対象物に照射した状態で検査対象物の画像を撮像装置にて撮像し、濃淡値を画素値とし且つ同一仕様の多数の検査対象物について同一部位が同一画素に対応する撮像画像を得る撮像過程を含み、前記多数の検査対象物を順次撮像して得られる撮像画像を用いて画像処理装置にて各検査対象物の欠陥の有無を検査する欠陥検査方法であって、検査対象物の撮像画像を設定用画像として設定用画像メモリに記憶する設定用記憶過程と、設定用画像メモリに蓄積された複数枚の設定用画像について画素ごとに濃淡値の度数分布を求める分布算出過程と、度数分布の結果を用い既定の設定ルールに従って濃淡値の良品範囲を設定する範囲設定過程と、撮像画像を撮像画像メモリに記憶する撮像記憶過程と、撮像画像の各画素ごとに濃淡値が前記良品範囲内にあるか否かを判断し、良品範囲内にある画素を良品画素、良品範囲内にない画素を不良画素と判定する画素判定過程と、撮像画像に含まれる不良画素の個数から検査対象物の欠陥の有無を判定する欠陥判定過程と、欠陥判定過程で欠陥がないと判定された場合に当該検査対象物の撮像画像を設定用画像として設定用記憶過程により設定用画像メモリに記憶し、分布算出過程および範囲設定過程により良品範囲を自動的に更新させる範囲更新過程とを有することを特徴とする欠陥検査方法。   An image of the inspection object is picked up by the image pickup device in a state where the inspection object is irradiated with light or X-rays, and the same part corresponds to the same pixel with respect to a large number of inspection objects having the same specification with grayscale values as pixel values. A defect inspection method that includes an imaging process for obtaining a captured image, and inspects each inspection object for defects using an image processing apparatus using captured images obtained by sequentially imaging the plurality of inspection objects; A setting storage process for storing a captured image of an inspection object as a setting image in a setting image memory, and a distribution for obtaining a frequency distribution of gray values for each pixel for a plurality of setting images stored in the setting image memory A calculation process, a range setting process for setting a non-defective range of gray values according to a predetermined setting rule using the result of the frequency distribution, an imaging storage process for storing the captured image in the captured image memory, and each pixel of the captured image A pixel determination process for determining whether or not a light value is within the non-defective range, determining a pixel within the non-defective range as a non-defective pixel, and determining a pixel not within the non-defective range as a defective pixel, and a defective pixel included in the captured image A defect determination process for determining the presence / absence of a defect in an inspection object from the number of inspection objects, and when it is determined that there is no defect in the defect determination process, a captured image of the inspection object is set as a setting image by a setting storage process A defect inspecting method comprising: a range update process which is stored in an image memory and automatically updates a non-defective range by a distribution calculation process and a range setting process. 前記画素判定過程で前記良品画素と判定された画素のうち前記不良画素と判定された画素で周囲を囲まれたものがある場合、良品画素と判定された当該画素については判定結果を不良画素と改める判定修正過程を有することを特徴とする請求項1記載の欠陥検査方法。   If there is a pixel surrounded by the pixels determined as the defective pixels among the pixels determined as the non-defective pixels in the pixel determination process, the determination result is determined as a defective pixel for the pixels determined as the non-defective pixels. 2. The defect inspection method according to claim 1, further comprising a determination correction process to be revised. 前記判定修正過程において判定結果が前記不良画素と改められた画素がある場合に、前記良品範囲について範囲を狭めるように修正する範囲修正過程を有することを特徴とする請求項2記載の欠陥検査方法。   3. The defect inspection method according to claim 2, further comprising a range correction process for correcting the non-defective range so as to narrow the range when there is a pixel whose determination result is changed to the defective pixel in the determination correction process. . 前記判定修正過程において判定結果が前記不良画素と改められた前記撮像画像が2枚以上の規定枚数に達した場合に、前記良品範囲について範囲を狭めるように修正する範囲修正過程を有することを特徴とする請求項2記載の欠陥検査方法。   And a range correction process for correcting the non-defective product range so that the range is narrowed when the number of captured images whose determination result is changed to the defective pixel in the determination correction process reaches a specified number of two or more. The defect inspection method according to claim 2. 前記設定用記憶過程では、前記設定用画像メモリに蓄積された前記設定用画像が設定用画像メモリに記憶可能な上限枚数に達すると、設定用画像メモリ内の設定用画像を古い側から順に削除しながら新しい設定用画像を記憶することを特徴とする請求項1ないし請求項4のいずれか1項に記載の欠陥検査方法。   In the setting storage process, when the number of setting images stored in the setting image memory reaches the maximum number that can be stored in the setting image memory, the setting images in the setting image memory are deleted in order from the oldest side. 5. The defect inspection method according to claim 1, wherein a new setting image is stored. 光またはX線を検査対象物に照射する照射装置と、検査対象物の画像を撮像して、濃淡値を画素値とし且つ同一仕様の多数の検査対象物について同一部位が同一画素に対応する撮像画像を得る撮像装置と、前記多数の検査対象物を順次撮像して得られる撮像画像を用いて各検査対象物の欠陥の有無を検査する画像処理装置とを備えた欠陥検査システムであって、画像処理装置は、検査対象物の撮像画像を設定用画像として記憶する設定用画像メモリと、設定用画像メモリに蓄積された複数枚の設定用画像について画素ごとに濃淡値の度数分布を求める分布算出手段と、度数分布の結果を用い既定の設定ルールに従って濃淡値の良品範囲を設定する範囲設定手段と、撮像画像を記憶する撮像画像メモリと、撮像画像の各画素ごとに濃淡値が前記良品範囲内にあるか否かを判断し、良品範囲内にある画素を良品画素、それ以外の画素を不良画素と判定する画素判定手段と、撮像画像に含まれる不良画素の個数から検査対象物の欠陥の有無を判定する欠陥判定手段と、欠陥判定手段で欠陥がないと判定された場合に当該検査対象物の撮像画像を設定用画像として設定用画像メモリに記憶し、分布算出手段および範囲設定手段により良品範囲を自動的に更新させる範囲更新手段とを有することを特徴とする欠陥検査システム。   An irradiation device that irradiates the inspection object with light or X-rays, and an image of the inspection object that is imaged so that the same value corresponds to the same pixel for a large number of inspection objects of the same specification with the gray value as the pixel value A defect inspection system comprising: an imaging device that obtains an image; and an image processing device that inspects for the presence or absence of defects in each inspection object using captured images obtained by sequentially imaging the numerous inspection objects, An image processing apparatus includes a setting image memory that stores a captured image of an inspection object as a setting image, and a distribution for obtaining a frequency distribution of gray values for each pixel for a plurality of setting images accumulated in the setting image memory. The calculation means, the range setting means for setting the non-defective range of the gray value according to the predetermined setting rule using the result of the frequency distribution, the captured image memory for storing the captured image, and the gray value for each pixel of the captured image An object to be inspected from the number of defective pixels included in a captured image, and pixel determination means for determining whether or not a pixel is in a non-defective range and determining that a pixel in the non-defective range is a non-defective pixel A defect determining means for determining the presence or absence of a defect, and when the defect determining means determines that there is no defect, a captured image of the inspection object is stored as a setting image in a setting image memory, and a distribution calculating means and range A defect inspection system comprising: a range update unit that automatically updates a non-defective product range by a setting unit.
JP2009077696A 2009-03-26 2009-03-26 Method for inspecting defect and defect inspection system Withdrawn JP2010230452A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2009077696A JP2010230452A (en) 2009-03-26 2009-03-26 Method for inspecting defect and defect inspection system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2009077696A JP2010230452A (en) 2009-03-26 2009-03-26 Method for inspecting defect and defect inspection system

Publications (1)

Publication Number Publication Date
JP2010230452A true JP2010230452A (en) 2010-10-14

Family

ID=43046434

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2009077696A Withdrawn JP2010230452A (en) 2009-03-26 2009-03-26 Method for inspecting defect and defect inspection system

Country Status (1)

Country Link
JP (1) JP2010230452A (en)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017012656A1 (en) * 2015-07-21 2017-01-26 Hewlett-Packard Development Company, L. P. Object generation temperature measurement
JP2017032548A (en) * 2015-06-26 2017-02-09 コグネックス・コーポレイション Using 3d vision for automated industrial inspection
WO2018088760A3 (en) * 2016-11-14 2018-08-16 주식회사 고영테크놀러지 Method and device for adjusting quality determination conditions for test body
JP2019029001A (en) * 2017-07-26 2019-02-21 キヤノン株式会社 Image data management method, manufacturing apparatus, production system, and image management method of production system
KR20190057221A (en) * 2019-05-16 2019-05-28 주식회사 고영테크놀러지 Method and apparatus for adjusting condition of quality decision for inspection target
KR20200013250A (en) * 2020-01-17 2020-02-06 주식회사 고영테크놀러지 Method and apparatus for adjusting condition of quality decision for inspection target
CN112014404A (en) * 2020-08-27 2020-12-01 Oppo(重庆)智能科技有限公司 Component detection method, device, system, electronic equipment and storage medium
JP2021014100A (en) * 2019-07-16 2021-02-12 Nok株式会社 Mold monitoring system
WO2021181793A1 (en) * 2020-03-13 2021-09-16 オムロン株式会社 Inspection system, inspection device, and inspection program
US11158039B2 (en) 2015-06-26 2021-10-26 Cognex Corporation Using 3D vision for automated industrial inspection
CN113740365A (en) * 2020-05-15 2021-12-03 东芝It·控制***株式会社 Nondestructive inspection device
US11366068B2 (en) 2016-11-14 2022-06-21 Koh Young Technology Inc. Inspection apparatus and operating method thereof
JP2022105581A (en) * 2018-03-08 2022-07-14 Ubeマシナリー株式会社 Method of detecting molding failure
US11836911B2 (en) 2017-07-26 2023-12-05 Canon Kabushiki Kaisha Image data management method, production apparatus, production system, image data management method for production system, and non-transitory computer-readable recording medium

Cited By (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2017032548A (en) * 2015-06-26 2017-02-09 コグネックス・コーポレイション Using 3d vision for automated industrial inspection
US11158039B2 (en) 2015-06-26 2021-10-26 Cognex Corporation Using 3D vision for automated industrial inspection
JP7037876B2 (en) 2015-06-26 2022-03-17 コグネックス・コーポレイション Use of 3D vision in automated industrial inspection
WO2017012656A1 (en) * 2015-07-21 2017-01-26 Hewlett-Packard Development Company, L. P. Object generation temperature measurement
CN107530961A (en) * 2015-07-21 2018-01-02 惠普发展公司有限责任合伙企业 Object generates temperature survey
US10539474B2 (en) 2015-07-21 2020-01-21 Hewlett-Packard Development Company, L.P. Object generation temperature measurement
CN107530961B (en) * 2015-07-21 2020-02-18 惠普发展公司有限责任合伙企业 Object-generated temperature measurement
WO2018088760A3 (en) * 2016-11-14 2018-08-16 주식회사 고영테크놀러지 Method and device for adjusting quality determination conditions for test body
US11199503B2 (en) 2016-11-14 2021-12-14 Koh Young Technology Inc. Method and device for adjusting quality determination conditions for test body
US11366068B2 (en) 2016-11-14 2022-06-21 Koh Young Technology Inc. Inspection apparatus and operating method thereof
US11836911B2 (en) 2017-07-26 2023-12-05 Canon Kabushiki Kaisha Image data management method, production apparatus, production system, image data management method for production system, and non-transitory computer-readable recording medium
JP7321676B2 (en) 2017-07-26 2023-08-07 キヤノン株式会社 Image data management method, manufacturing apparatus, production system, image management method for production system, control program, and recording medium
JP2019029001A (en) * 2017-07-26 2019-02-21 キヤノン株式会社 Image data management method, manufacturing apparatus, production system, and image management method of production system
JP2022105581A (en) * 2018-03-08 2022-07-14 Ubeマシナリー株式会社 Method of detecting molding failure
KR102070858B1 (en) 2019-05-16 2020-01-29 주식회사 고영테크놀러지 Method and apparatus for adjusting condition of quality decision for inspection target
KR20190057221A (en) * 2019-05-16 2019-05-28 주식회사 고영테크놀러지 Method and apparatus for adjusting condition of quality decision for inspection target
JP2021014100A (en) * 2019-07-16 2021-02-12 Nok株式会社 Mold monitoring system
KR102167558B1 (en) 2020-01-17 2020-10-19 주식회사 고영테크놀러지 Method and apparatus for adjusting condition of quality decision for inspection target
KR20200013250A (en) * 2020-01-17 2020-02-06 주식회사 고영테크놀러지 Method and apparatus for adjusting condition of quality decision for inspection target
WO2021181793A1 (en) * 2020-03-13 2021-09-16 オムロン株式会社 Inspection system, inspection device, and inspection program
CN113740365A (en) * 2020-05-15 2021-12-03 东芝It·控制***株式会社 Nondestructive inspection device
CN112014404A (en) * 2020-08-27 2020-12-01 Oppo(重庆)智能科技有限公司 Component detection method, device, system, electronic equipment and storage medium

Similar Documents

Publication Publication Date Title
JP2010230452A (en) Method for inspecting defect and defect inspection system
JP5260183B2 (en) Defect inspection method and apparatus
US20140119638A1 (en) System, method and computer program product to evaluate a semiconductor wafer fabrication process
JP2006200972A (en) Image defect inspection method, image defect inspection device, and external appearance inspection device
JP5556346B2 (en) Wafer defect inspection apparatus and wafer defect inspection method
US7561751B2 (en) Image processing method
JP2008180696A (en) Defect detector, defect detecting method, image sensor device, image sensor module, defect detecting program, and computer readable recording medium
KR20200100086A (en) Systems and methods for setting up production line inspection
JP5196572B2 (en) Wafer storage cassette inspection apparatus and method
CN108445010B (en) Automatic optical detection method and device
JPWO2013128616A1 (en) Method and apparatus for detecting display unevenness of display device
KR20140091916A (en) Inspection Method For Display Panel
US10274314B2 (en) Shape inspection method, shape inspection apparatus, and program
CN115298539A (en) Appearance inspection system and computer program
KR20160118722A (en) 3D vision inpection system
JP2009229197A (en) Linear defect detecting method and device
JP6349804B2 (en) Inspection device, inspection method, program, and recording medium thereof
JP2005121546A (en) Defect inspection method
JP4244046B2 (en) Image processing method and image processing apparatus
JP2005165387A (en) Method and device for detecting stripe defective of picture and display device
JP4743231B2 (en) Appearance inspection method and appearance inspection apparatus
JP2013117490A (en) Inspection system and method for setting recipe
KR20110020437A (en) Inspection method of patterned wafer for detecting defects
JP6114559B2 (en) Automatic unevenness detector for flat panel display
JP2006275691A (en) Inspection method and inspection device

Legal Events

Date Code Title Description
RD04 Notification of resignation of power of attorney

Free format text: JAPANESE INTERMEDIATE CODE: A7424

Effective date: 20100714

A300 Withdrawal of application because of no request for examination

Free format text: JAPANESE INTERMEDIATE CODE: A300

Effective date: 20120605