JPH06123706A - Material component analyzer - Google Patents
Material component analyzerInfo
- Publication number
- JPH06123706A JPH06123706A JP27125292A JP27125292A JPH06123706A JP H06123706 A JPH06123706 A JP H06123706A JP 27125292 A JP27125292 A JP 27125292A JP 27125292 A JP27125292 A JP 27125292A JP H06123706 A JPH06123706 A JP H06123706A
- Authority
- JP
- Japan
- Prior art keywords
- spark
- image
- image information
- sample
- input
- 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.)
- Pending
Links
Landscapes
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
Description
【０００１】[0001]
【産業上の利用分野】この発明は火花試験手段を使用し
た材料（例えば鋼材）の成分判定装置に関するものであ
る。BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a material (e.g., steel) component determination device using a spark test means.
【０００２】[0002]
【従来の技術】鋼材の成分判定手段には鋼材のサンプル
を切り出し、成分分析を行うことで精密な測定ができ
る。この場合における分析機器も測定技術も確立してい
て一般に使用されている。しかし、種々の材料が流れて
いる生産現場で、鋼材の成分を簡便に判定するために、
現在、火花試験による手段や火花を発生させる手応えを
用いて行う手段が実施されている。例えば火花試験手段
としては日本工業規格ＪＩＳ Ｇ ０５６６で制定され
ている。2. Description of the Related Art Precise measurement can be performed by cutting out a sample of a steel material as a means for determining the composition of a steel material and analyzing the composition. In this case, analytical instruments and measurement techniques have been established and are commonly used. However, in production sites where various materials are flowing, in order to easily determine the composition of steel materials,
At present, a means for performing a spark test and a means for using a response to generate a spark are being implemented. For example, the spark test means is established in Japanese Industrial Standard JIS G 0566.
【０００３】上記試験は火花を観察して、火花の状態か
らいかなる鋼種であるかを推定したり、異材混入の鑑別
や異材が混入していないかを判定するために行われる。
火花を発生させる手段としては図３に示すように、グラ
インダ１に試料である鋼材２を接触させて行う。火花が
図３のように生じたとき、その火花状態をグラインダ１
からＡ1までを根木、Ａ1からＡ2までを中央、Ａ2以降を
先端と称している。また、火花が飛んでいる線分を流線
と称し、火花が飛散している部分を破裂と称している。The above-mentioned test is carried out by observing sparks to estimate what kind of steel it is from the state of the sparks, to discriminate whether or not foreign materials are mixed, and to judge whether or not foreign materials are mixed.
As a means for generating a spark, as shown in FIG. 3, a steel material 2 as a sample is brought into contact with a grinder 1 to carry out a spark. When a spark is generated as shown in Fig. 3, change the spark condition to the grinder 1
Leek up A 1 from is called from A 1 to A 2 center, a tip A 2 or later. A line segment in which sparks are flying is called a streamline, and a part in which sparks are flying is called a burst.
【０００４】[0004]
【発明が解決しようとする課題】前述のように鋼材の成
分判定には２つの手段がある。前者の鋼材のサンプルを
切り出して成分分析を行う方法は、サンプリ切り出しか
ら分析までの工程、作業時間、設備コストの面から流れ
作業工程（インライン）に使用することができない不具
合がある。As described above, there are two means for determining the composition of the steel material. The former method of cutting out a sample of a steel material and performing component analysis has a problem that it cannot be used in a flow work process (in-line) in terms of steps from cutting out a sample to analysis, work time, and equipment cost.
【０００５】また、後者の火花判定方法では熟練者の目
視による手段を採っているため、熟練者の判定能力に負
わざるを得なく、自動判定がいまだにできない問題があ
る。Further, in the latter method of determining a spark, since a means for visually checking by a skilled person is employed, there is no choice but to bear the ability of the skilled person to make a determination, and there is a problem that automatic determination is still impossible.
【０００６】この発明は上記の事情に鑑みてなされたも
ので、流れ作業工程に使用できるとともに、熟練者に頼
ることなく簡単に材料の成分が判別できる材料成分判定
装置を提供することを目的とする。The present invention has been made in view of the above circumstances, and an object thereof is to provide a material component judging device which can be used in a flow work process and can easily judge the constituents of a material without relying on a skilled person. To do.
【０００７】[0007]
【課題を解決するための手段】この発明は上記の目的を
達成するために、第１発明は試料から火花を発生させる
火花発生部と、この火花発生部からの火花の画像を入力
する画像入力部と、この画像入力部に入力された火花の
画像情報を取り込み、ニューラルネットワークを用い
て、その画像情報から火花の特徴量を学習及び推定して
出力する特徴抽出部と、前記画像入力部に入力された火
花の画像情報を取り込み、その画像情報から火花の複雑
さをフラクタル次元演算を用いて出力する演算部と、こ
の演算部からの火花の複雑さと前記特徴抽出部からの火
花の特徴量とが入力され、両関係をファジィ推論して前
記試料の組成情報を得る総合判断部とからなることを特
徴とするものである。SUMMARY OF THE INVENTION In order to achieve the above object, the present invention provides a spark generating part for generating a spark from a sample, and an image input for inputting an image of the spark from the spark generating part. Part, a feature extraction part that captures image information of the spark input to the image input part, learns and estimates the feature amount of the spark from the image information by using a neural network, and outputs the feature amount, and An arithmetic unit that captures the image information of the input spark and outputs the complexity of the spark from the image information by using the fractal dimension calculation, and the complexity of the spark from this arithmetic unit and the feature amount of the spark from the feature extraction unit. Is inputted, and a fuzzy inference of both relations is performed to obtain the composition information of the sample, and a comprehensive judgment unit is provided.
【０００８】第２発明は実際の火花と総合判断部に与え
られた成分把握手段による画像情報とから試料の組成情
報を得るようにしたものである。The second aspect of the invention is to obtain the composition information of the sample from the actual spark and the image information by the component grasping means provided to the comprehensive judgment section.
【０００９】第３発明は種々の試料が流れている作業工
程に火花発生部を設けたことを特徴とするものである。A third aspect of the invention is characterized in that a spark generating portion is provided in a work process in which various samples are flowing.
【００１０】[0010]
【作用】試料から発生する火花をＣＣＤカメラ等による
画像入力部に取り込む。取り込んだ画像は電気信号とし
て出力し、ニューラルネットワークを用いて火花の画像
パターンの特徴量を学習及び推定するとともに前記電気
信号をフラクタル次元演算を行って火花の複雑さを計算
する。火花の特徴量と複雑さとの関係をファジィ推論を
用いて組成情報を得る。[Operation] Sparks generated from the sample are taken into the image input section such as a CCD camera. The captured image is output as an electric signal, and the feature amount of the image pattern of the spark is learned and estimated using a neural network, and the complexity of the spark is calculated by performing the fractal dimension calculation of the electric signal. The composition information is obtained by using fuzzy reasoning about the relation between the feature quantity of sparks and complexity.
【００１１】[0011]
【実施例】以下この発明の実施例を図面に基づいて説明
する。図１において、１１は図３に示すグランダ等を用
いて試料から火花を発生させる火花発生部で、この火花
発生部１１で発生された火花の画像はＣＣＤカメラ等か
ら構成される画像入力部１２に入力される。画像入力部
１２は入力された火花の画像をＲ（赤），Ｇ（緑），Ｂ
（青）三原色に分解した画像情報Ｐを出力する。画像情
報Ｐはニューラルネットワークを用いて火花の画像パタ
ーンの特徴量を学習及び推定する特徴量抽出部１３に入
力され、出力に学習された火花画像の特徴に対する画像
情報Ｐの類似度情報Ｎを得る。Embodiments of the present invention will be described below with reference to the drawings. In FIG. 1, reference numeral 11 denotes a spark generation unit for generating a spark from a sample using the grounder shown in FIG. 3, and the image of the spark generated by the spark generation unit 11 is an image input unit 12 including a CCD camera or the like. Entered in. The image input unit 12 converts the input spark image into R (red), G (green), and B.
The image information P separated into (blue) three primary colors is output. The image information P is input to the feature amount extraction unit 13 that learns and estimates the feature amount of the spark image pattern using the neural network, and the similarity information N of the image information P with respect to the learned feature of the spark image is obtained at the output. .
【００１２】また、前記画像情報Ｐはフラクタル次元演
算部１３に入力され、ここで画像情報Ｐから火花の複雑
さを計算して出力にフラクタル次元情報Ｆを得る。フラ
クタル次元演算部１３からの出力情報（火花の複雑さ）
Ｆと前記特徴量抽出部１３からの出力情報（火花画像の
特徴）Ｎはファジィ推論からなる総合判断部１５に入力
され、ここで、両情報の関係をファジィ推論して出力に
試料の組成情報Ｃを得る。なお、上記実施例において、
火花発生部１１はアーク等の手段で発生させてもよい。Further, the image information P is input to the fractal dimension calculation unit 13, where the complexity of the spark is calculated from the image information P and the fractal dimension information F is obtained at the output. Output information from fractal dimension calculator 13 (spark complexity)
F and the output information (feature of the spark image) N from the feature amount extraction unit 13 are input to the comprehensive determination unit 15 which is a fuzzy inference, where the fuzzy inference is performed on the relationship between both information and the composition information of the sample is output. Get C. In the above embodiment,
The spark generation part 11 may be generated by means such as an arc.
【００１３】図２は図１の実施例の動作モードを述べる
ためのブロック図で、この図２の動作モードは学習モー
ドと組成判定モードがある。学習モードは特徴抽出部１
３であるニューラルネットワークのパラメータを、画像
情報Ｐに対する教師信号Ｔ（代表的な材料番号など）を
用いて学習させて出力Ｎ1を得る。この出力Ｎ1とフラク
タル次元演算部１３の火花の複雑さ出力情報Ｆは総合判
断部１５に入力され、図示しない発光分析装置等による
分析結果Ｒを用いてファジィ推論で用いるファジィ推論
ルール及びメンバーシップ関数を自動生成する。FIG. 2 is a block diagram for explaining the operation mode of the embodiment shown in FIG. 1. The operation mode of FIG. 2 includes a learning mode and a composition judging mode. The learning mode is the feature extraction unit 1
The parameter of the neural network, which is 3, is learned by using the teacher signal T (representative material number or the like) for the image information P to obtain the output N 1 . This output N 1 and the spark complexity output information F of the fractal dimension calculation unit 13 are input to the comprehensive judgment unit 15, and the fuzzy inference rules and membership used in the fuzzy inference using the analysis result R by an optical emission analyzer (not shown) or the like. Function is automatically generated.
【００１４】一方、組成判別モードは上記学習モードに
より学習した装置に実際の火花画像を取り込み、取り込
んだ画像情報から試料の組成を推定するモードである。On the other hand, the composition discriminating mode is a mode in which an actual spark image is captured by the apparatus learned in the learning mode and the composition of the sample is estimated from the captured image information.
【００１５】上記両モードを用いることにより、実際の
火花と他の発光分析法等による成分把握の画像情報によ
って火花の特徴量とその複雑さ及び成分とを関連づける
学習を行い、実作動時に未知の試料（鋼材）の火花を観
測して組成を判別することができるようになる。By using both of the above modes, learning is performed by associating the characteristic amount of the spark with the complexity and the component based on the image information of the actual spark and the component grasping by another emission analysis method, etc. It becomes possible to determine the composition by observing the spark of the sample (steel material).
【００１６】上記実施例を作業工程中に設置すれば、作
動工程に流れている試料（鋼種）と異なった材質（鋼
材）が混入したことが判別できる。これによって、異材
混入に伴う莫大な被害（出荷後、部品に混入し、組み込
まれた製品の事故の発生、ロット不良に伴う損失、信頼
性失墜等）を未然に防止して製品の信頼性を確立でき
る。また、鋼材の製造工程中の前後ロットへの混入防止
を図ることができるとともに、特殊鋼等リサイクルのた
めの成分管理を行って異成分の混入防止を図ることがで
きる。例えば、ステンレスのＮｉ系、Ｃｒ系に異材を混
入させないことや、鋼のリサイクルにＣｕやＺｎを混入
させないこと等である。If the above embodiment is installed during the working process, it can be determined that a material (steel material) different from the sample (steel type) flowing in the working process is mixed. This prevents the enormous damage caused by mixing different materials (such as the occurrence of accidents in products that have been mixed into parts after shipment and has been incorporated, loss due to defective lots, loss of reliability, etc.), and improves product reliability. Can be established. In addition, it is possible to prevent mixing of the steel material into the front and rear lots during the manufacturing process, and it is possible to prevent mixing of different components by controlling the composition of the special steel for recycling. For example, it is necessary not to mix different materials into stainless steel such as Ni and Cr, and to prevent mixing of Cu and Zn into steel recycling.
【００１７】上記の他、機械部品は切断→切削→熱処理
→矯正→研摩する工程、切断→鍛造→バリ取り→切削→
熱処理等多工程、機械工程や熱処理工程等種々な工程を
経て来るので、途中の工程で部品に異材が混入するのを
防止できる。In addition to the above, mechanical parts are cut → cut → heat treated → straightened → abraded, cutting → forged → deburred → cut →
Since various processes such as heat treatment, multi-process, mechanical process and heat treatment process are performed, it is possible to prevent foreign materials from being mixed into parts in the middle of the process.
【００１８】[0018]
【発明の効果】以上述べたように、この発明によれば、
流れ作業工程中に配置して試料と異なる材質の材料の混
入を防止することができるとともに、熟練者でなくとも
簡単に試料の成分を判別することができる利点がある。As described above, according to the present invention,
There is an advantage that it can be arranged during the flow operation process to prevent mixing of a material of a material different from that of the sample, and that even an unskilled person can easily determine the components of the sample.
【図１】この発明の実施例を示すブロック図である。FIG. 1 is a block diagram showing an embodiment of the present invention.
【図２】図１の動作モードを述べるブロック図である。FIG. 2 is a block diagram illustrating an operation mode of FIG.
【図３】火花発生部における火花の形および名称を説明
図である。FIG. 3 is an explanatory diagram of a shape and a name of a spark in a spark generation unit.
１１…火花発生部 １２…画像入力部 １３…特徴抽出部 １４…フラクタル次元演算部 １５…総合判断部 11 ... Spark generation part 12 ... Image input part 13 ... Feature extraction part 14 ... Fractal dimension calculation part 15 ... Comprehensive judgment part
Claims (3)
と、この火花発生部からの火花の画像を入力する画像入
力部と、この画像入力部に入力された火花の画像情報を
取り込み、ニューラルネットワークを用いて、その画像
情報から火花の特徴量を学習及び推定して出力する特徴
抽出部と、前記画像入力部に入力された火花の画像情報
を取り込み、その画像情報から火花の複雑さをフラクタ
ル次元演算を用いて出力する演算部と、この演算部から
の火花の複雑さと前記特徴抽出部からの火花の特徴量と
が入力され、両関係をファジィ推論して前記試料の組成
情報を得る総合判断部とからなることを特徴とする材料
成分判定装置。1. A neural network for generating a spark from a sample, an image input section for inputting an image of the spark from the spark generation section, and image information of the spark input to the image input section to obtain a neural network. Using a feature extraction unit that learns and estimates the feature amount of the spark from the image information and outputs the feature amount, the image information of the spark input to the image input unit is taken in, and the complexity of the spark is fractal from the image information. A calculation unit that outputs using a dimensional calculation, the complexity of the spark from this calculation unit and the feature amount of the spark from the feature extraction unit are input, and fuzzy inference is performed on both relationships to obtain the composition information of the sample. A material component determination device comprising: a determination unit.
分把握手段による画像情報とから試料の組成情報を得る
請求項１記載の材料成分判定装置。2. The material component determination apparatus according to claim 1, wherein composition information of the sample is obtained from actual sparks and image information obtained by the component grasping means provided to the comprehensive determination unit.
発生部を設けたことを特徴とする請求項１記載の材料成
分判定装置。3. The material component determination device according to claim 1, wherein a spark generation part is provided in a work process in which various samples are flowing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP27125292A JPH06123706A (en) | 1992-10-09 | 1992-10-09 | Material component analyzer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP27125292A JPH06123706A (en) | 1992-10-09 | 1992-10-09 | Material component analyzer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| JPH06123706A true JPH06123706A (en) | 1994-05-06 |
Family
ID=17497486
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP27125292A Pending JPH06123706A (en) | 1992-10-09 | 1992-10-09 | Material component analyzer |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPH06123706A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6208981B1 (en) * | 1995-07-26 | 2001-03-27 | Siemens Aktiengesellschaft | Circuit configuration for controlling a running-gear or drive system in a motor vehicle |
| WO2010004947A1 (en) * | 2008-07-08 | 2010-01-14 | 住友金属工業株式会社 | Device for determining quality of steel material and method for determining quality of steel material |
-
1992
- 1992-10-09 JP JP27125292A patent/JPH06123706A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6208981B1 (en) * | 1995-07-26 | 2001-03-27 | Siemens Aktiengesellschaft | Circuit configuration for controlling a running-gear or drive system in a motor vehicle |
| WO2010004947A1 (en) * | 2008-07-08 | 2010-01-14 | 住友金属工業株式会社 | Device for determining quality of steel material and method for determining quality of steel material |
| JP4958025B2 (en) * | 2008-07-08 | 2012-06-20 | 住友金属工業株式会社 | Steel material judgment device and steel material judgment method |
| US8498445B2 (en) | 2008-07-08 | 2013-07-30 | Nippon Steel & Sumitomo Metal Corporation | Material determining apparatus for steel product and material determining method for steel product |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US5751910A (en) | Neural network solder paste inspection system | |
| JP3172153B2 (en) | Oil pollution degree measuring device | |
| JP2968442B2 (en) | Evaluation system for welding defects | |
| US5218555A (en) | Method for judging a color difference using rules fuzzy inference and apparatus therefor | |
| AU2021259109A1 (en) | Method, device and computer program for determining the performance of a welding method via digital processing of an image of the welded workpiece | |
| CN113780900B (en) | Welding detection system and method based on edge calculation | |
| Spitz et al. | Inspection of sheet materials—test of model predictions | |
| Dobrzański et al. | Computer aided classification of flaws occurred during casting of aluminum | |
| Cook et al. | Automated visual inspection and interpretation system for weld quality evaluation | |
| JPH06123706A (en) | Material component analyzer | |
| Chen | Computer vision for general purpose visual inspection: a fuzzy logic approach | |
| CN1668920A (en) | Method for online measurement of molten phases | |
| CN112718710B (en) | Method for implementing laser cleaning of copper substrate pollutants based on red, green and blue numerical analysis | |
| JPH05337663A (en) | Welding state monitor | |
| JPH09120455A (en) | Feature discriminating method using neural network | |
| US5517237A (en) | Video photometric color system for processing color specific streams | |
| JP2022013285A (en) | Machine learning method | |
| Bidanda et al. | Development of a spatter index for automated welding inspection using computer vision | |
| CN118609041A (en) | High-power laser-arc composite welding process defect identification method and system | |
| CN116921817B (en) | Automatic TIG welding arc concentration on-line monitoring and intelligent early warning method | |
| Rummel | Human factors considerations in the assessment of nondestructive evaluation (NDE) reliability | |
| Klinieam et al. | A Machine Vision Approach for Copper Plating Quality Control System Development | |
| JPH0915138A (en) | Method for judging quality of steel or the like and spark control tool for the judging | |
| Mauris et al. | Fuzzy techniques evaluate sausage quality | |
| Scheideler et al. | In-Situ Process Monitoring in Additive Manufacturing Using Statistics and Pre-Process Data |