JPH07280746A - Apparatus for extracting flaw from metal plate surface - Google Patents

Apparatus for extracting flaw from metal plate surface

Info

Publication number
JPH07280746A
JPH07280746A JP6071897A JP7189794A JPH07280746A JP H07280746 A JPH07280746 A JP H07280746A JP 6071897 A JP6071897 A JP 6071897A JP 7189794 A JP7189794 A JP 7189794A JP H07280746 A JPH07280746 A JP H07280746A
Authority
JP
Japan
Prior art keywords
metal plate
image
pixel
plate surface
value
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Withdrawn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
JP6071897A
Other languages
Japanese (ja)
Inventor
Riyouichi Danki
亮一 段木
Akira Miyajima
明 宮嶌
Tetsuji Uetake
徹司 植竹
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.)
JFE Steel Corp
Original Assignee
Kawasaki Steel Corp
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 Kawasaki Steel Corp filed Critical Kawasaki Steel Corp
Priority to JP6071897A priority Critical patent/JPH07280746A/en
Publication of JPH07280746A publication Critical patent/JPH07280746A/en
Withdrawn legal-status Critical Current

Links

Landscapes

  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

PURPOSE:To extract flaws from a metal plate surface with high precision by installing a line computing means to obtain a peak line or a dale line which passes picture element of interest and is formed by ranging picture elements with the maximum luminance value or the minimum luminance value. CONSTITUTION:A light and shade image sent from an image pick-up element 1 is converted by an A/D converting part 2 and saved in an image data memory part 3. The coordinate of a picture element with the maximum picture element value is detected out of the light and shade image of the memory part 3 by comparing the picture element values one another, which are converted into digital values, by a comparison part 4. The detected coordinate and the luminance value are saved in an observing picture element memory part 5. Line computing means 6a, 6b compute a peak line or a dale line which passes the observing picture element and is formed by vertically ranging the picture elements with the maximum luminance value or the minimum luminance value in respectively specified transverse direction of the metal plate. Based on the pattern of respective picture elements, cut point detecting means 8a, 8b detect the cut point which cross the edge of a crack. The region containing the flaws surrounded with the edge coordinates saved in a memory part 9 is cut by a cutting part 10.

Description

【発明の詳細な説明】Detailed Description of the Invention

【0001】[0001]

【産業上の利用分野】本発明は、金属板表面画像を表わ
す画像データに基づいて、金属板表面の疵を抽出する金
属板表面疵抽出装置に関する。
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a metal plate surface flaw extracting device for extracting flaws on the surface of a metal plate based on image data representing a surface image of the metal plate.

【0002】[0002]

【従来の技術】例えば、鋼板製造工程で発生した鋼板表
面の疵を高精度に抽出することが求められている。従
来、鋼板等の金属板の表面疵を検出する場合、二値化手
法が用いられ、撮像素子で取り込んだ画像全体を、ある
1つのしきい値を決定してそのしきい値で二値化するの
が一般的である(例えば特開平4−286083号公報
参照)。
2. Description of the Related Art For example, it is required to accurately extract a flaw on a surface of a steel sheet that has occurred in a steel sheet manufacturing process. Conventionally, in the case of detecting a surface flaw of a metal plate such as a steel plate, a binarization method is used, and a certain one threshold value is determined for the entire image captured by the image pickup device and the threshold value is binarized. This is generally done (see, for example, Japanese Patent Application Laid-Open No. 4-286083).

【0003】図6は、この二値化の様子を示した図であ
り、図6(a)は、撮像素子から入力された濃淡画像の
一部、図6(b)は、図6(a)に示す部分の画素値の
プロファイル、図6(c),(d),(e)は、それぞ
れ、図6(b)に示す各しきい値A,B,Cで二値化さ
れたときの二値画像を示している。画像全体に対し同一
のしきい値を適用すると撮像素子から入力された濃淡画
像が二値画像に変換されたとき、もとの濃淡画像の形状
情報が失われる。これは濃淡画像が通常の撮像素子で撮
像したときは、例えば256階調の濃度情報をもってい
るのに対し、二値画像の濃淡情報が2階調であることか
ら生じる。つまり、256階調の濃淡情報を2階調の濃
淡情報に変換するとき情報が失われ、図形の形状情報が
失われるからである。
FIG. 6 is a diagram showing the state of this binarization. FIG. 6 (a) is a part of a grayscale image input from an image sensor, and FIG. 6 (b) is FIG. 6 (a). 6 (c), 6 (d), and 6 (e) are the pixel value profiles of the portion shown in FIG. 6) when binarized with the thresholds A, B, and C shown in FIG. 6 (b), respectively. 2 shows a binary image of. If the same threshold value is applied to the entire image, the shape information of the original grayscale image is lost when the grayscale image input from the image sensor is converted into a binary image. This occurs because the grayscale image has density information of, for example, 256 gradations when the grayscale image is picked up by a normal image pickup device, whereas the grayscale information of the binary image has two gradations. That is, the information is lost when converting the grayscale information of 256 gradations into the grayscale information of 2 gradations, and the shape information of the figure is also lost.

【0004】この様な情報欠落を防ぐため、特開平4−
33176号公報には、ある程度の領域ごとにしきい値
を求め1つの画像に対しいくつものしきい値を設けるこ
とが提案されている。入力された画像を各水平走査線お
よび各垂直走査線ごとに画素値の極大値および極小値を
求め、隣り合った極大値と極小値の値から次々と二値化
のしきい値を求めていく。これを適用した例を図7を用
いて説明する。
In order to prevent such information loss, Japanese Unexamined Patent Publication No.
Japanese Laid-Open Patent Publication No. 33176 proposes to obtain a threshold value for each region to some extent and to provide a number of threshold values for one image. The maximum and minimum values of the pixel value of the input image are calculated for each horizontal scanning line and each vertical scanning line, and the binarization threshold value is calculated one after another from the adjacent maximum and minimum values. Go. An example in which this is applied will be described with reference to FIG.

【0005】図7において、(a)は撮像素子から入力
された濃淡画像の一部、(b)の(b1)は(a)の部
分の画素値のプロファイル、(c)は(b2)を各領域
ごとのしきい値としたときの二値画像を示している。図
7の例からもわかる様に、各領域ごとにしきい値を設定
した二値化処理においては濃淡画像の形状情報が失われ
ずに所望の二値画像が得られる。
In FIG. 7, (a) is a part of a grayscale image input from the image pickup device, (b) of (b1) is a pixel value profile of the (a) part, and (c) is (b2). The binary image when the threshold value is set for each area is shown. As can be seen from the example of FIG. 7, in the binarization process in which the threshold value is set for each area, the desired binary image can be obtained without losing the shape information of the grayscale image.

【0006】しかしながら、これらの方法は画像全体に
着目しているために不必要な画素まで抽出してしまい、
金属板表面の疵が存在する濃淡画像から疵部のみを高精
度に抽出することは不可能である。
However, since these methods focus on the entire image, unnecessary pixels are extracted,
It is impossible to extract only the flawed portion with high accuracy from the grayscale image in which the flaw on the surface of the metal plate is present.

【0007】[0007]

【発明が解決しようとする課題】本発明は、上記事情に
鑑み、金属板表面の疵を高精度に抽出することのできる
金属板表面疵抽出装置を提供することを目的とする。
SUMMARY OF THE INVENTION In view of the above circumstances, it is an object of the present invention to provide a metal plate surface flaw extraction device capable of extracting a flaw on the surface of a metal plate with high accuracy.

【0008】[0008]

【課題を解決するための手段】上記目的を達成する本発
明の金属板表面疵抽出装置は、 (1)金属板表面画像を表わす画像データを記憶する画
像データ記憶部 (2)画像データに基づいて、画像中の金属板表面疵を
含むとともに該疵の外周に沿って一巡する画像領域を切
り出す領域切出部 (3)上記画像領域の中から金属板表面疵の存在する画
像領域を抽出する疵抽出部 を備えたことを特徴とする。
MEANS FOR SOLVING THE PROBLEMS A metal plate surface flaw extracting device of the present invention which achieves the above object is (1) an image data storage unit for storing image data representing a metal plate surface image. (2) based on image data. Area cut-out portion that includes a metal plate surface flaw in the image and cuts out an image area that makes a round along the outer circumference of the flaw (3) Extract an image area in which the metal plate surface flaw exists from the image area It is characterized by having a flaw extraction unit.

【0009】ここで上記本発明の金属板表面疵抽出装置
において、前記領域切出部は、 (2−1)前記画像中の、画素値の最大値および最小値
のうち、金属板表面疵としてあらわれる側の一方の値を
有する注目画素もしくはその一方の値を含みその一方の
値に近い値を有する複数の注目画素を検出する注目画素
検出手段 (2−2)注目画素を通り、金属板の所定の各横方向の
極大輝度値あるいは極小輝度の画素を縦方向に連ねた峰
線もしくは谷線を求める線演算手段 (2−3)上記峰線もしくは谷線上の各画素から、上記
縦方向と交わる横方向に、その横方向に並ぶ各画素をサ
ーチし、各画素の画素値のパターンに基づいて、金属板
表面疵のエッジを越えた切出し点を検出する切出し点検
出手段 (2−4)上記切出し点により囲まれた画像領域を切り
出す切出し手段 を備えたものであることが好ましい。
Here, in the above-described metal plate surface flaw extraction device of the present invention, the area cutout portion is (2-1) as a metal plate surface flaw among the maximum and minimum pixel values in the image. Target pixel detecting means for detecting a target pixel having one value on the appearing side or a plurality of target pixels including one value and having a value close to the one value (2-2) Passing through the target pixel, a target plate Line calculation means for obtaining a peak line or a valley line in which pixels having a predetermined maximum luminance value or a minimum luminance value in each horizontal direction are connected in the vertical direction (2-3) From each pixel on the peak line or the valley line to the vertical direction Cutout point detection means for searching each pixel arranged in the horizontal direction in the intersecting horizontal direction and detecting a cutout point beyond the edge of the metal plate surface flaw based on the pixel value pattern of each pixel (2-4) Image surrounded by the above clipping points It is preferable to have a cutting-out means for cutting out a region.

【0010】また、上記本発明の金属板表面疵抽出装置
において、上記疵抽出部は、上記(2)の領域切出部で
切り出された画像領域について二値化処理を行なうこと
により、金属板表面疵の存在する画像領域を抽出する二
値化手段を備えたものであることが好ましい。本発明の
金属板表面疵抽出装置は、金属板表面疵を含むとともに
その疵の外周に沿って一巡する画像領域を切り出し、そ
の画像領域のみについて例えば二値化手法を適用して金
属板表面疵を抽出するものであるため、不必要な画素ま
で抽出される可能性が低減され、その金属板表面疵が高
精度に抽出される。
In the metal plate surface flaw extracting apparatus of the present invention, the flaw extracting unit performs binarization processing on the image area cut out by the area cutting section in (2) above to obtain a metal plate. It is preferable to include a binarizing means for extracting an image region having surface flaws. The metal plate surface flaw extraction device of the present invention is a metal plate surface flaw that includes a metal plate surface flaw and cuts out an image region that makes a round along the outer periphery of the flaw, and applies, for example, a binarization method only to the image region. Is extracted, the possibility that even unnecessary pixels are extracted is reduced, and the metal plate surface flaw is extracted with high accuracy.

【0011】また本発明の金属板表面疵抽出装置におい
て、上記(2)の領域切出部が上記(2−1)〜(2−
4)の各手段を備えたものである場合、疵を含む画像領
域が十分な精度で抽出される。
Further, in the metal plate surface flaw extracting apparatus of the present invention, the area cutout portion of the above (2) is (2-1) to (2-).
When each means of 4) is provided, the image area including the flaw is extracted with sufficient accuracy.

【0012】[0012]

【実施例】以下、本発明の実施例について説明する。こ
こでは本発明で対象としている金属板の一例である鋼板
表面の疵を抽出する例について説明する。図1は、鋼板
表面の画像を取り込む撮像系の配置を示した図である。
鋼板が図示の矢印L方向に搬送されている。光源は、約
30°傾いた方向、鋼板から約300mm離れた位置に
配置されており、その鋼板表面が約2万luxの照度で
照射されている。また、ラインセンサを内蔵したカメラ
が、光源から照射された光の正反射光が入射しないよう
に、鋼板垂直方向約950mm離れた位置に配置され、
そのカメラにより、鋼板表面の暗視野の画像が取り込ま
れる。
EXAMPLES Examples of the present invention will be described below. Here, an example of extracting a flaw on the surface of a steel plate, which is an example of a metal plate targeted by the present invention, will be described. FIG. 1 is a diagram showing the arrangement of an imaging system for capturing an image of the surface of a steel plate.
The steel sheet is conveyed in the direction of the arrow L shown in the figure. The light source is arranged in a direction inclined by about 30 ° and at a position apart from the steel plate by about 300 mm, and the surface of the steel plate is illuminated with an illuminance of about 20,000 lux. In addition, the camera with the built-in line sensor is arranged at a position approximately 950 mm away from the vertical direction of the steel plate so that specularly reflected light emitted from the light source does not enter.
The camera captures a dark field image of the steel plate surface.

【0013】このとき、鋼板表面疵は、取り込まれた画
像上では画素値の大きい領域としてあらわれる。図2
は、本発明の金属板表面疵抽出装置の一実施例の構成を
示すブロック図である。撮像素子1から入力された濃淡
画像はA/D変換部2により濃淡値がアナログ値からデ
ジタル値に変換され、画像データ記憶部3にデジタル化
された濃淡画像が記憶される。デジタル値に変換された
画素値を比較部4で各々比較することによって、画像デ
ータ記憶部3の濃淡画像から最も画素値の大きい画素座
標を検出する。検出された座標とその濃度値は注目画素
記憶部5に記憶される。本実施例では、比較部4が、本
発明にいう注目画素検出手段の一例である。
At this time, the steel plate surface flaw appears as an area having a large pixel value on the captured image. Figure 2
FIG. 1 is a block diagram showing a configuration of an embodiment of a metal plate surface flaw extraction device of the present invention. The grayscale image input from the image sensor 1 is converted from an analog value to a digital value by the A / D conversion unit 2, and the digitized grayscale image is stored in the image data storage unit 3. The pixel values converted into digital values are compared by the comparison unit 4, and the pixel coordinates having the largest pixel value are detected from the grayscale image in the image data storage unit 3. The detected coordinates and their density values are stored in the target pixel storage unit 5. In the present embodiment, the comparison unit 4 is an example of the pixel-of-interest detection unit referred to in the present invention.

【0014】注目画素記憶部5の注目画素の座標を
(A,B)とすると、その上(図1に示すL方向に1つ
ずれた位置)に並ぶ3画素(A−1,B−1),(A,
B−1),(A+1,B−1)の画素値を比較部6aで
比較し最も画素値の大きかった画素を次の注目画素とし
て、さらにその上3画素の画素値を比較部6aで比較し
最も画素値の大きかった画素を次の注目画素するという
ようにこれを画像の最上部まで行う。同様に、注目画素
の下3画素の濃度値を比較部6bで比較し最も濃度値の
大きかった画素の座標が次の注目画素というようにこれ
を画像の最下部まで行う。比較部6a,6bで注目画素
とされた座標は尾根線座標記憶部7に記憶される。本実
施例では、比較部6a,6bが、本発明にいう線演算手
段の一例である。
Assuming that the coordinates of the target pixel in the target pixel storage unit 5 are (A, B), the three pixels (A-1, B-1) arranged on the position (a position shifted by one in the L direction shown in FIG. 1) are located above it. ), (A,
The pixel values of (B-1) and (A + 1, B-1) are compared by the comparison unit 6a, the pixel having the largest pixel value is set as the next pixel of interest, and the pixel values of three pixels are further compared by the comparison unit 6a. Then, this is performed up to the top of the image such that the pixel having the largest pixel value is set as the next pixel of interest. Similarly, the density values of the three lower pixels of the target pixel are compared by the comparison unit 6b, and the coordinates of the pixel having the highest density value are the next target pixel, and this is performed up to the bottom of the image. The coordinates selected as the target pixel by the comparison units 6a and 6b are stored in the ridge line coordinate storage unit 7. In this embodiment, the comparison units 6a and 6b are an example of the line calculation means according to the present invention.

【0015】図3は、この尾根線を求める手法を図解し
た模式図である。撮像素子等から入力された画像の中で
濃度値が最も大きな画素座標を(A,B)とし、この注
目画素を中心としてy軸の上下方向それぞれに以下の座
標移動を行なう。図3に示すように上方向の座標移動は
注目画素の上3画素の座標(A−1,B−1),(A,
B−1),(A+1,B−1)の中で最も画素値の大き
い画素を次の注目画素とする移動である。この注目画素
に対しても同様の座標移動を行なう。これを画像の一番
上まで繰り返す。ただし、注目画素の上3画素のうち、
最も画素値の大きい画素が右か左の画素の場合、さらに
その隣の画素を参照し、その画素の画素値の方がさらに
大きい場合はさらにその隣の画素を参照するというよう
にして画素値が極大となる画素を見つけ、その画素を注
目画素とする。下方向の座標移動に対しても上方向と同
様に画像の一番下まで行なう。これらの注目画素をプロ
ットしていくと一本の線になる。これを画素値の尾根線
と呼ぶことにする。
FIG. 3 is a schematic diagram illustrating a method for obtaining this ridge line. The pixel coordinate having the largest density value in the image input from the image sensor or the like is defined as (A, B), and the following coordinate movement is performed in the up-down direction of the y-axis with this pixel of interest as the center. As shown in FIG. 3, the coordinate movement in the upward direction is caused by the coordinates (A-1, B-1), (A,
This is a movement in which the pixel having the largest pixel value among B-1) and (A + 1, B-1) is set as the next target pixel. The same coordinate movement is performed for this target pixel. Repeat this to the top of the image. However, of the top three pixels of the pixel of interest,
When the pixel with the largest pixel value is the right or left pixel, the pixel next to it is referenced, and when the pixel value of that pixel is larger, the pixel next to it is referenced. Finds the pixel at which is the maximum, and sets that pixel as the pixel of interest. The downward coordinate movement is performed to the bottom of the image as in the upward direction. Plotting these pixels of interest results in a single line. This will be called a ridge line of pixel values.

【0016】図4は、直線y=B上の画素値プロファイ
ルを示す図である。図2に示す右エッジ判定部8aにお
いて、図4に示すように、尾根線座標記憶部7の座標を
探索開始点としてその右の画素の画素値を順次抽出して
いき画素値が次に上がるところまでの間で最も画素値が
平坦になる点、すなわち、一次徴分が0を横切る点の画
素ないしその点に最も近い画素を右のエッジ座標とす
る。同様に左エッジ判定部8bにおいて、尾根線座標記
憶部7の座標を探索開始点としてその左の画素の画素値
を順次抽出していき画素値が平坦になる点の画素を左の
エッジ座標とする。本実施例では、左右のエッジ判定部
8a,8bが、本発明にいう切出し点検出手段の一例で
ある。
FIG. 4 is a diagram showing a pixel value profile on the straight line y = B. In the right edge determination unit 8a shown in FIG. 2, as shown in FIG. 4, the coordinates of the ridge line coordinate storage unit 7 are used as the search start point to sequentially extract the pixel value of the pixel on the right, and the pixel value increases next. The point at which the pixel value is most flat up to that point, that is, the pixel at the point where the primary characteristic crosses 0 or the pixel closest to that point is the right edge coordinate. Similarly, in the left edge determination unit 8b, the pixel value of the left pixel is sequentially extracted with the coordinates of the ridge line coordinate storage unit 7 as the search start point, and the pixel at the point where the pixel value becomes flat is defined as the left edge coordinate. To do. In the present embodiment, the left and right edge determination sections 8a and 8b are an example of the cut-out point detection means according to the present invention.

【0017】左右のエッジ判定部8a,8bより左右の
エッジ座標の検出を上下方向それぞれに順次行ってい
き、左右のエッジ座標が同一になった場合エッジ検出を
終了し、そこまでの左右のエッジ座標がエッジ座標記憶
部9に記憶される。図5は、エッジ座標記憶部9に記憶
されたエッジ座標の、画像上の位置を示す模式図であ
る。
The left and right edge determination units 8a and 8b sequentially detect the left and right edge coordinates in the vertical direction. When the left and right edge coordinates are the same, the edge detection is ended and the left and right edges up to that point are detected. The coordinates are stored in the edge coordinate storage unit 9. FIG. 5 is a schematic diagram showing positions on the image of the edge coordinates stored in the edge coordinate storage unit 9.

【0018】エッジ座標記憶部9に記憶されたエッジ座
標で囲まれた領域が切り出し部10で切り出され、疵を
含む画像領域が求められる。本実施例では、この切り出
し部10が、本発明にいう切出し手段の一例である。こ
のようにして切り出された画像領域は、次に、二値化部
11に入力される。二値化部11では、例えば画像全体
から定めたしきい値等による単純な二値化手法を用いて
もよいが、好ましくは判別分析法二値化により疵の領域
が抽出される。
An area surrounded by the edge coordinates stored in the edge coordinate storage section 9 is cut out by the cutout section 10 to obtain an image area including a flaw. In the present embodiment, the cutout section 10 is an example of the cutout means according to the present invention. The image area cut out in this way is then input to the binarization unit 11. The binarization unit 11 may use a simple binarization method using, for example, a threshold value determined from the entire image, but preferably, the defect area is extracted by the discriminant analysis method binarization.

【0019】判別分析法二値化とは、あるしきい値によ
って濃度ヒストグラムを2クラスに分割した場合のクラ
ス間分散σB 2(k)が最大になるしきい値を選び、その
しきい値を用いて二値化する手法をいい、具体的には以
下の演算による。濃淡レベル(画素値)を{1,2,
…,L}とし、画素をレベル[1,k]と、レベル[k
+1,L]との2クラスに分割するしきい値をしきい値
kとする。このとき、クラス間分散σB 2は、 σB 2(k)=ω0 (μ0 −μT2 +ω1 (μ1 −μ
T2 ただし ni =濃度iの画素数 N =全画素数 pi =ni /N
Discriminant analysis method binarization means selecting a threshold value that maximizes the interclass variance σ B 2 (k) when the density histogram is divided into two classes by a certain threshold value, and the threshold value is selected. It is a method of binarization using, and is specifically calculated by the following calculation. The gray level (pixel value) is {1, 2,
, L}, the pixel is level [1, k], and level [k
The threshold value for dividing into two classes, such as +1, L], is set as a threshold value k. At this time, the interclass variance σ B 2 is σ B 2 (k) = ω 00 −μ T ) 2 + ω 11 −μ
T ) 2 where n i = number of pixels of density i N = total number of pixels p i = n i / N

【0020】[0020]

【数1】 [Equation 1]

【0021】で表わされ、このクラス間分散σB 2(k)
が最大となるしきい値kが選択される。本実施例では、
以上のようにして鋼板表面疵が高精度に抽出される。
And the interclass variance σ B 2 (k)
The threshold value k that maximizes is selected. In this embodiment,
As described above, the steel plate surface flaws are extracted with high accuracy.

【0022】[0022]

【発明の効果】以上説明したように、本発明の金属板表
面疵抽出装置は、金属板表面疵を含みその疵を一巡する
画像領域を切り出し、その画像領域から疵を抽出するも
のであるため、余計な画素が疵として認識される可能性
が低く、疵が高精度に抽出される。
As described above, since the metal plate surface flaw extraction device of the present invention is for extracting an image area including a metal plate surface flaw and surrounding the flaw, and extracting the flaw from the image area. , It is unlikely that extra pixels are recognized as flaws, and flaws are extracted with high accuracy.

【図面の簡単な説明】[Brief description of drawings]

【図1】鋼板表面の画像を取り込む撮像系の配置を示し
た図である。
FIG. 1 is a diagram showing an arrangement of an image pickup system for capturing an image of a steel plate surface.

【図2】本発明の金属板表面疵抽出装置の一実施例の構
成を示すブロック図である。
FIG. 2 is a block diagram showing the configuration of an embodiment of the metal plate surface flaw extraction device of the present invention.

【図3】尾根線を求める手法を図解した模式図である。FIG. 3 is a schematic diagram illustrating a method for obtaining a ridge line.

【図4】直線y=B上の画素値プロファイルを示す模式
図である。
FIG. 4 is a schematic diagram showing a pixel value profile on a straight line y = B.

【図5】エッジ座標記憶部9に記憶されたエッジ座標
の、画像上の位置を示す模式図である。
FIG. 5 is a schematic diagram showing a position on an image of edge coordinates stored in an edge coordinate storage unit 9.

【図6】二値化の様子を示した図である。FIG. 6 is a diagram showing how binarization is performed.

【図7】変化するしきい値を用いた二値化の様子を示し
た図である。
FIG. 7 is a diagram showing how binarization is performed using changing threshold values.

【符号の説明】[Explanation of symbols]

1 撮像素子 2 A/D変換部 4 比較部(注目画素検出手段) 6a,6b 比較部(線演算手段) 8a,8b エッジ判定部(切出し点検出手段) 10 切り出し部(切出し手段) 15 二値化部(二値化手段) DESCRIPTION OF SYMBOLS 1 Image sensor 2 A / D conversion unit 4 Comparison unit (target pixel detection unit) 6a, 6b Comparison unit (line calculation unit) 8a, 8b Edge determination unit (cutout point detection unit) 10 Cutout unit (cutout unit) 15 Binary Binization unit (binarization means)

Claims (3)

【特許請求の範囲】[Claims] 【請求項1】 金属板表面画像を表わす画像データを記
憶する画像データ記憶部と、 前記画像データに基づいて、前記画像中の金属板表面疵
を含むとともに該疵の外周に沿って一巡する画像領域を
切り出す領域切出部と、 前記画像領域の中から金属板表面疵の存在する画像領域
を抽出する疵抽出部とを備えたことを特徴とする金属板
表面疵抽出装置。
1. An image data storage unit that stores image data representing a surface image of a metal plate, and an image that includes a metal plate surface flaw in the image and that makes a round along the outer periphery of the flaw based on the image data. A metal plate surface flaw extraction device comprising: a region cutout portion that cuts out a region; and a flaw extraction portion that extracts an image region in which a metal plate surface flaw exists from the image region.
【請求項2】 前記領域切出部が、 前記画像中の、画素値の最大値および最小値のうち、金
属板表面疵としてあらわれる側の一方の値を有する注目
画素もしくは該一方の値を含み該一方の値に近い値を有
する複数の注目画素を検出する注目画素検出手段と、 該注目画素を通り、金属板の所定の各横方向の極大輝度
値あるいは極小輝度値の画素を縦方向に連ねた峰線もし
くは谷線を求める線演算手段と、 前記峰線もしくは谷線上の各画素から、前記縦方向と交
わる横方向に、該横方向に並ぶ各画素をサーチし、該各
画素の画素値のパターンに基づいて、金属板表面疵のエ
ッジを越えた切出し点を検出する切出し点検出手段と、 前記切出し点により囲まれた画像領域を切り出す切出し
手段とを備えたことを特徴とする請求項1記載の金属板
表面疵抽出装置。
2. The area cutout portion includes a pixel of interest or a value of one of the maximum value and the minimum value of pixel values in the image, which has one of the values that appears as a metal plate surface flaw. A target pixel detection unit that detects a plurality of target pixels having a value close to the one value, and a pixel having a predetermined maximum or minimum brightness value in each horizontal direction of the metal plate passing through the target pixel in the vertical direction. Line calculation means for obtaining a continuous peak line or valley line, and from each pixel on the peak line or valley line, in the horizontal direction intersecting the vertical direction, search for each pixel arranged in the horizontal direction, and the pixel of each pixel A cutout point detecting means for detecting a cutout point beyond the edge of the metal plate surface flaw based on the value pattern, and a cutout means for cutting out an image region surrounded by the cutout point are provided. Item 1 metal plate surface Extraction device.
【請求項3】 前記疵抽出部が、前記領域切出部で切り
出された画像領域について二値化処理を行なうことによ
り、金属板表面疵の存在する画像領域を抽出する二値化
手段を備えたことを特徴とする請求項1記載の金属板表
面疵抽出装置。
3. The flaw extraction unit includes a binarization unit that extracts an image region having a metal plate surface flaw by performing a binarization process on the image region cut out by the region cutout unit. The metal plate surface flaw extraction device according to claim 1, characterized in that.
JP6071897A 1994-04-11 1994-04-11 Apparatus for extracting flaw from metal plate surface Withdrawn JPH07280746A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP6071897A JPH07280746A (en) 1994-04-11 1994-04-11 Apparatus for extracting flaw from metal plate surface

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP6071897A JPH07280746A (en) 1994-04-11 1994-04-11 Apparatus for extracting flaw from metal plate surface

Publications (1)

Publication Number Publication Date
JPH07280746A true JPH07280746A (en) 1995-10-27

Family

ID=13473792

Family Applications (1)

Application Number Title Priority Date Filing Date
JP6071897A Withdrawn JPH07280746A (en) 1994-04-11 1994-04-11 Apparatus for extracting flaw from metal plate surface

Country Status (1)

Country Link
JP (1) JPH07280746A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0880023A1 (en) * 1997-05-23 1998-11-25 Siemag Transplan Gmbh Method and device for the automatic detection of surface faults during the continuous mechanical removal of material from casted products
JP2000193420A (en) * 1998-10-19 2000-07-14 Toyota Central Res & Dev Lab Inc Eye position detecting device
CN107817246A (en) * 2016-09-02 2018-03-20 富士通株式会社 Medium, image processing method and the image processing apparatus of storage image processing routine
CN117237646A (en) * 2023-11-15 2023-12-15 深圳市润海电子有限公司 PET high-temperature flame-retardant adhesive tape flaw extraction method and system based on image segmentation

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0880023A1 (en) * 1997-05-23 1998-11-25 Siemag Transplan Gmbh Method and device for the automatic detection of surface faults during the continuous mechanical removal of material from casted products
AU743622B2 (en) * 1997-05-23 2002-01-31 Siemag Transplan Gmbh Method and device for the automatic detection of surface defects for continuously cast products with continuous mechanical removal of the material
KR100536118B1 (en) * 1997-05-23 2006-02-28 지마크 트란스플란 게엠베하 Method and device for the automatic detection of surface defects for continuously cast products with continuous mechanical removal of the material
JP2000193420A (en) * 1998-10-19 2000-07-14 Toyota Central Res & Dev Lab Inc Eye position detecting device
CN107817246A (en) * 2016-09-02 2018-03-20 富士通株式会社 Medium, image processing method and the image processing apparatus of storage image processing routine
CN107817246B (en) * 2016-09-02 2020-06-19 富士通株式会社 Medium storing image processing program, image processing method and image processing apparatus
CN117237646A (en) * 2023-11-15 2023-12-15 深圳市润海电子有限公司 PET high-temperature flame-retardant adhesive tape flaw extraction method and system based on image segmentation
CN117237646B (en) * 2023-11-15 2024-01-30 深圳市润海电子有限公司 PET high-temperature flame-retardant adhesive tape flaw extraction method and system based on image segmentation

Similar Documents

Publication Publication Date Title
EP1269394B1 (en) Improved method for image binarization
JPH06203154A (en) Image processor
JPH0737103A (en) Inclination angle detection device
JP2006292593A (en) Flaw detector, flaw detecting method, computer program and recording medium
KR20070005968A (en) Method and apparatus for detecting defects of the welded part
JP2017521011A (en) Symbol optical detection method
JP3890844B2 (en) Appearance inspection method
JPH07280746A (en) Apparatus for extracting flaw from metal plate surface
JPH06207909A (en) Inspection apparatus for surface defect
JPH08305795A (en) Character recognizing method
JP4293653B2 (en) Appearance inspection method
JP2000028539A (en) Defect detecting device
JP2001021332A (en) Surface inspecting device and its method
JP3230509B2 (en) Moving image processing device
JP4403036B2 (en) Soot detection method and apparatus
JP2005346222A (en) Sweat gland pore removing device, sweat gland pore removing method and sweat gland pore removing program
JPH0718812B2 (en) Foreign object detection method
JP2000357287A (en) Method and device for number plate recognition
JP2004286708A (en) Defect detection apparatus, method, and program
JPH0688789A (en) Method for extracting density fluctuation constitution picture element in image and method for judging concentration fluctuation mass
JP4544955B2 (en) Image monitoring device
JP3038092B2 (en) Appearance inspection method
JPH0797410B2 (en) Image processing method
JPH0815172A (en) Visual inspection method
JPH10214327A (en) Method of detecting surface defect using image

Legal Events

Date Code Title Description
A300 Withdrawal of application because of no request for examination

Free format text: JAPANESE INTERMEDIATE CODE: A300

Effective date: 20010703