JP6759785B2 - Liquid level shape extraction method, equipment and program - Google Patents

Liquid level shape extraction method, equipment and program Download PDF

Info

Publication number
JP6759785B2
JP6759785B2 JP2016137867A JP2016137867A JP6759785B2 JP 6759785 B2 JP6759785 B2 JP 6759785B2 JP 2016137867 A JP2016137867 A JP 2016137867A JP 2016137867 A JP2016137867 A JP 2016137867A JP 6759785 B2 JP6759785 B2 JP 6759785B2
Authority
JP
Japan
Prior art keywords
liquid level
time
difference
original data
horizontal direction
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.)
Active
Application number
JP2016137867A
Other languages
Japanese (ja)
Other versions
JP2018009847A (en
Inventor
山本 浩貴
浩貴 山本
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.)
Nippon Steel Corp
Original Assignee
Nippon 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 Nippon Steel Corp filed Critical Nippon Steel Corp
Priority to JP2016137867A priority Critical patent/JP6759785B2/en
Publication of JP2018009847A publication Critical patent/JP2018009847A/en
Application granted granted Critical
Publication of JP6759785B2 publication Critical patent/JP6759785B2/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Measurement Of Levels Of Liquids Or Fluent Solid Materials (AREA)
  • Continuous Casting (AREA)
  • Image Analysis (AREA)

Description

本発明は、容器内の自由液面を撮影した、時間的に連続する画像に基づいて液面形状を抽出する液面形状の抽出方法、装置及びプログラムに関する。 The present invention relates to a liquid level shape extraction method, an apparatus and a program for extracting a liquid level shape based on a time-continuous image of a free liquid level in a container.

鉄鋼製造プロセスにおける鋼の連続鋳造工程においては、鋳型内の湯面挙動が鋳片欠陥に影響することが知られている。
従来から、鋳型内の溶鋼の流動状況、湯面挙動を適正化するのに必要な操業因子を明らかにするために、水モデル実験が用いられている(非特許文献1を参照)。水モデル実験とは、透明なアクリル樹脂板等を用いて鋳型を模した容器を作成し、溶鋼の代わりに水を満たすことで、溶鋼の流動状況、湯面挙動を模擬する実験である。このような水モデル実験において水面高さを定量測定する必要がある場合には、非特許文献1で示されるように、通常、フロートとレーザー変位計を利用し、レーザー変位計の設置位置での水面高さを測定することが行われるが、この場合、レーザー変位計で測定していない位置での水面高さを知ることはできなかった。
In the continuous casting process of steel in the steel manufacturing process, it is known that the behavior of the molten metal in the mold affects the slab defects.
Conventionally, a water model experiment has been used in order to clarify the flow state of molten steel in a mold and the operating factors necessary for optimizing the molten metal surface behavior (see Non-Patent Document 1). The water model experiment is an experiment in which a container imitating a mold is created using a transparent acrylic resin plate or the like and filled with water instead of molten steel to simulate the flow state and molten metal surface behavior of the molten steel. When it is necessary to quantitatively measure the water surface height in such a water model experiment, as shown in Non-Patent Document 1, a float and a laser displacement meter are usually used, and the laser displacement meter is installed at the installation position. The water surface height is measured, but in this case, it was not possible to know the water surface height at a position not measured by the laser displacement meter.

このような課題に対し、特許文献1では、液体が注水される水槽を挟んで、一方の側に光源を設置し、他方の側に近赤外CCDカメラを設置して、近赤外CCDカメラの撮影画像から輝度比(水槽が空の位置と注水位置との輝度比)を算出し、エッジ処理により液面位置を検出する手法が開示されている。
また、非特許文献2では、ビデオカメラ等の撮影画像から、画像処理技術により曲線を検出する手法が開示されている。非特許文献2では、微分フィルタで山岳の稜線を含むエッジを強調した上で、稜線の連続性を利用し動的計画法により稜線を抽出する手法が提案されている。
In response to such a problem, in Patent Document 1, a light source is installed on one side and a near-infrared CCD camera is installed on the other side of a water tank into which a liquid is injected. A method of calculating the brightness ratio (brightness ratio between the position where the water tank is empty and the water injection position) and detecting the liquid level position by edge processing is disclosed.
Further, Non-Patent Document 2 discloses a method of detecting a curve by an image processing technique from an image taken by a video camera or the like. Non-Patent Document 2 proposes a method of emphasizing an edge including a mountain ridge with a differential filter and then extracting the ridge by a dynamic programming method using the continuity of the ridge.

特開2005−291830号公報Japanese Unexamined Patent Publication No. 2005-291830

手嶋ら:スラブ高速鋳造時の連鋳鋳型内溶鋼流動におよぼす鋳造条件の影響,鉄と鋼,vol.79,No.5,1992.Teshima et al .: Effect of casting conditions on molten steel flow in continuous casting mold during high-speed slab casting, Iron and Steel, vol.79, No.5, 1992. W. Lie, T. T. Lin, K. Hung, A robust dynamic programming algorithm to extract skyline in images for navigation. Pattern. Recogn. Lett. 26, 221-230 (2005).W. Lie, T. T. Lin, K. Hung, A robust dynamic programming algorithm to extract skyline in images for navigation. Pattern. Recogn. Lett. 26, 221-230 (2005). R.A.Maronna,R.D.Martin,V.J.Yohai:Robust Statistics:Theory and Methods,Wiley,2006.R.A.Maronna, R.D.Martin, V.J.Yohai: Robust Statistics: Theory and Methods, Wiley, 2006.

しかしながら、水のような無色透明な液体の気液界面付近は、輝度のコントラストが小さく、特許文献1のような単純なエッジ処理では、画像中のノイズ成分と気液界面を見分けることは困難で、気液界面を見誤ることがある。
また、非特許文献2の手法では、動的計画法による曲線(液面)の抽出の際、画像中のノイズに反応し、液面を誤識別することがある。
このように気液界面を見誤ったり、液面を誤識別したりすると、正確な液体の流動状況、液面挙動を把握できなくなってしまう。
However, the contrast of brightness is small near the gas-liquid interface of a colorless and transparent liquid such as water, and it is difficult to distinguish the noise component in the image from the gas-liquid interface by simple edge processing as in Patent Document 1. , The gas-liquid interface may be misunderstood.
Further, in the method of Non-Patent Document 2, when extracting a curve (liquid level) by a dynamic programming method, the liquid level may be erroneously identified in response to noise in an image.
If the gas-liquid interface is misunderstood or the liquid level is misidentified in this way, it becomes impossible to accurately grasp the liquid flow state and the liquid level behavior.

本発明は上記のような点に鑑みてなされたものであり、容器内の自由液面を撮影した、時間的に連続する画像に基づいて、正確な液体の流動状況、液面挙動を把握できるようにすることを目的とする。 The present invention has been made in view of the above points, and an accurate liquid flow state and liquid level behavior can be grasped based on time-continuous images of the free liquid level in the container. The purpose is to do so.

上記の課題を解決するための本発明の要旨は、以下のとおりである。
[1] 容器内の自由液面を撮影した、時間的に連続する複数の画像に基づいて液面形状を抽出する液面形状の抽出方法であって、
前記時間的に連続する画像を用いて作成される、各時間において一水平方向の各位置で液面高さが単一に定まる原データを対象として、
前記原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項とを含む評価関数に基づいて、前記一水平方向の各位置、各時刻の液面高さを推定し、
前記原データと液面高さの計算値との差分に関する項は、jを前記一水平方向の各位置(j=1、2、・・・、M)、tを各時刻(t=1、2、・・・、T)とし、前記原データにおける液面高さy j (t)と、液面高さの計算値d j (t)との誤差絶対値|y j (t)−d j (t)|の前記一水平方向の各位置、各時刻に関する和で表わされる、
ことを特徴とする液面形状の抽出方法。
] スラック変数を用いて前記誤差絶対値を不等式制約に置き換え、前記評価関数を2次計画法により解くことを特徴とする[]に記載の液面形状の抽出方法。
] 前記液面高さの計算値の時間差分に関する項は、時間方向における液面高さの1階差分又は2階差分の二乗和で表わされることを特徴とする[1]又は[2]に記載の液面形状の抽出方法。
] 前記液面高さの計算値の空間差分に関する項は、空間方向における液面高さの1階差分又は2階差分の二乗和で表わされることを特徴とする[1]乃至[]のいずれか一つに記載の液面形状の抽出方法。
] 容器内の自由液面を撮影した、時間的に連続する複数の画像に基づいて液面形状を抽出する液面形状の抽出装置であって、
前記時間的に連続する画像を用いて作成される、各時間において一水平方向の各位置で液面高さが単一に定まる原データを対象として、
前記原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項とを含む評価関数に基づいて、前記一水平方向の各位置、各時刻の液面高さを推定する推定手段を備え
前記原データと液面高さの計算値との差分に関する項は、jを前記一水平方向の各位置(j=1、2、・・・、M)、tを各時刻(t=1、2、・・・、T)とし、前記原データにおける液面高さy j (t)と、液面高さの計算値d j (t)との誤差絶対値|y j (t)−d j (t)|の前記一水平方向の各位置、各時刻に関する和で表わされる、
ことを特徴とする液面形状の抽出装置。
] 容器内の自由液面を撮影した、時間的に連続する複数の画像に基づいて液面形状を抽出するためのプログラムであって、
前記時間的に連続する画像を用いて作成される、各時間において一水平方向の各位置で液面高さが単一に定まる原データを対象として、
前記原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項とを含む評価関数に基づいて、前記一水平方向の各位置、各時刻の液面高さを推定し、
前記原データと液面高さの計算値との差分に関する項は、jを前記一水平方向の各位置(j=1、2、・・・、M)、tを各時刻(t=1、2、・・・、T)とし、前記原データにおける液面高さy j (t)と、液面高さの計算値d j (t)との誤差絶対値|y j (t)−d j (t)|の前記一水平方向の各位置、各時刻に関する和で表わされる、処理をコンピュータに実行させるためのプログラム。
The gist of the present invention for solving the above problems is as follows.
[1] A liquid level shape extraction method for extracting a liquid level shape based on a plurality of images taken of a free liquid level in a container, which are continuous in time.
For the original data in which the liquid level height is fixed at each position in one horizontal direction at each time, which is created using the temporally continuous images.
An evaluation function that includes a term relating to the difference between the original data and the calculated value of the liquid level, a term relating to the time difference of the calculated value of the liquid level, and a term relating to the spatial difference of the calculated value of the liquid level. Based on this, the liquid level at each position in the one horizontal direction and at each time is estimated .
Regarding the term regarding the difference between the original data and the calculated value of the liquid level, j is each position (j = 1, 2, ..., M) in the one horizontal direction, and t is each time (t = 1, 2, ..., T), and the absolute value of the error between the liquid level height y j (t) in the original data and the calculated liquid level height d j (t) | y j (t) −d It is represented by the sum of each position and each time in the one horizontal direction of j (t) |.
A method for extracting the liquid level shape.
[ 2 ] The method for extracting a liquid level shape according to [ 1 ], wherein the absolute value of the error is replaced with an inequality constraint by using a slack variable, and the evaluation function is solved by a quadratic programming method.
[ 3 ] The term relating to the time difference of the calculated value of the liquid level height is represented by the sum of squares of the first-order difference or the second-order difference of the liquid level height in the time direction [1] or [2 ] . ] . The method for extracting the liquid level shape.
[ 4 ] The term relating to the spatial difference of the calculated liquid level height is represented by the sum of squares of the first-order difference or the second-order difference of the liquid level in the spatial direction [1] to [ 3 ]. ]. The method for extracting the liquid level shape according to any one of the above.
[ 5 ] A liquid level shape extraction device that extracts the liquid level shape based on a plurality of temporally continuous images of the free liquid level in the container.
For the original data in which the liquid level height is fixed at each position in one horizontal direction at each time, which is created using the temporally continuous images.
An evaluation function that includes a term relating to the difference between the original data and the calculated value of the liquid level, a term relating to the time difference of the calculated value of the liquid level, and a term relating to the spatial difference of the calculated value of the liquid level. Based on the above, an estimation means for estimating the liquid level at each position and each time in the one horizontal direction is provided .
Regarding the term regarding the difference between the original data and the calculated value of the liquid level, j is each position (j = 1, 2, ..., M) in the one horizontal direction, and t is each time (t = 1, 2, ..., T), and the absolute value of the error between the liquid level height y j (t) in the original data and the calculated liquid level height d j (t) | y j (t) −d It is represented by the sum of each position and each time in the one horizontal direction of j (t) |.
A liquid level shape extraction device characterized by this.
[ 6 ] A program for extracting the liquid level shape based on a plurality of temporally continuous images of the free liquid level in the container.
For the original data in which the liquid level height is fixed at each position in one horizontal direction at each time, which is created using the temporally continuous images.
An evaluation function that includes a term relating to the difference between the original data and the calculated value of the liquid level, a term relating to the time difference of the calculated value of the liquid level, and a term relating to the spatial difference of the calculated value of the liquid level. Based on this, the liquid level at each position in the one horizontal direction and at each time is estimated .
Regarding the term regarding the difference between the original data and the calculated value of the liquid level, j is each position (j = 1, 2, ..., M) in the one horizontal direction, and t is each time (t = 1, 2, ..., T), and the absolute value of the error between the liquid level height y j (t) in the original data and the calculated liquid level height d j (t) | y j (t) −d A program for causing a computer to execute a process represented by the sum of each position and each time in the one horizontal direction of j (t) | .

本発明によれば、原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項とを含む評価関数に基づいて、一水平方向の各位置、各時刻の液面高さを推定するようにしたので、原データにおいて液面の誤識別が生じていても、その誤識別を取り除いて、時間方向及び空間方向になめらかに変動する液面形状を抽出することができる。これにより、正確な液体の流動状況、液面挙動を把握することが可能となる。 According to the present invention, there are a term relating to the difference between the original data and the calculated value of the liquid level, a term relating to the time difference of the calculated value of the liquid level, and a term relating to the spatial difference of the calculated value of the liquid level. Since the liquid level height at each position and time in one horizontal direction is estimated based on the evaluation function including, even if the liquid level is misidentified in the original data, the misidentification is removed. , It is possible to extract a liquid level shape that fluctuates smoothly in the temporal direction and the spatial direction. This makes it possible to accurately grasp the flow state and liquid level behavior of the liquid.

実施形態に係る液面形状の抽出装置の機能構成を示す図である。It is a figure which shows the functional structure of the liquid level shape extraction apparatus which concerns on embodiment. 実施形態に係る液面形状の抽出装置による液面形状の抽出方法を示すフローチャートである。It is a flowchart which shows the extraction method of the liquid level shape by the liquid level shape extraction apparatus which concerns on embodiment. 原データと、原データを対象として液面形状を抽出した結果とを示す特性図である。It is a characteristic diagram which shows the original data and the result of extracting the liquid level shape with respect to the original data. 原データにおける液面形状を表わす写真と、原データを対象として抽出した液面形状を表わす写真を示す図である。It is a figure which shows the photograph which shows the liquid level shape in the original data, and the photograph which shows the liquid level shape extracted for the original data. 本発明を適用して液面形状を抽出した結果の例を表わす写真を示す図である。It is a figure which shows the photograph which shows the example of the result of extracting the liquid surface shape by applying this invention.

以下、添付図面を参照して、本発明の好適な実施形態について説明する。
図1に、実施形態に係る液面形状の抽出装置100の機能構成を示す。
図1に示すように、透明なアクリル樹脂板等を用いて鋳型を模した容器200を作成し、溶鋼の代わりに水201を満たす。そして、容器200の一面に対向させるようにしてカメラ300を設置し、カメラ300により、容器200内の自由液面を撮影した時間的に連続する画像、ここでは動画像を取得する。カメラ300で撮影する画像は、可視光画像に限らず、赤外線画像等としてもよい。
Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.
FIG. 1 shows the functional configuration of the liquid level shape extraction device 100 according to the embodiment.
As shown in FIG. 1, a container 200 imitating a mold is prepared using a transparent acrylic resin plate or the like, and water 201 is filled in place of molten steel. Then, the camera 300 is installed so as to face one surface of the container 200, and the camera 300 acquires a temporally continuous image of the free liquid surface in the container 200, in this case, a moving image. The image captured by the camera 300 is not limited to a visible light image, but may be an infrared image or the like.

液面形状の抽出装置100において、101は入力部であり、カメラ300で撮影した容器200内の自由液面の動画像データを入力する。 In the liquid level shape extraction device 100, 101 is an input unit, and inputs moving image data of the free liquid level in the container 200 taken by the camera 300.

102は原データ作成部であり、入力部101で入力する動画像データを用いて、各時間において一水平方向(以下、幅方向と呼ぶ)の各位置で液面高さが単一に定まる原データを作成する。原データは、各時刻での画像について、例えば高さ方向に微分フィルタ演算を行い、その輝度変化のピーク位置を液面と判定する手法で作成される。或いは、原データは、各時刻での画像について、例えば非特許文献2のように動的計画法により曲線を抽出する手法で作成される。 Reference numeral 102 denotes an original data creation unit, which uses the moving image data input by the input unit 101 to determine the liquid level at each position in the horizontal direction (hereinafter referred to as the width direction) at each time. Create data. The original data is created by a method of performing a differential filter calculation on the image at each time, for example, in the height direction, and determining the peak position of the brightness change as the liquid level. Alternatively, the original data is created for the image at each time by a method of extracting a curve by a dynamic programming method as in Non-Patent Document 2, for example.

103は液面高さ推定部であり、原データ作成部102で作成した原データを対象として、原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項とを含む評価関数を構築し、この評価関数を最小にする幅方向の各位置、各時刻の液面高さを推定する。 Reference numeral 103 denotes a liquid level height estimation unit, which targets the original data created by the original data creation unit 102, and includes a term relating to the difference between the original data and the calculated liquid level height and the calculated liquid level height. Build an evaluation function that includes a term related to the time difference and a term related to the spatial difference of the calculated liquid level height, and estimate each position in the width direction that minimizes this evaluation function and the liquid level height at each time. ..

104は出力部であり、液面高さ推定部103で推定した液面高さから得られる液面形状の抽出結果を出力する。例えば液面形状の抽出結果を表示装置に表示したり、ネットワークを介して外部機器に送信したりする。 Reference numeral 104 denotes an output unit, which outputs the extraction result of the liquid level shape obtained from the liquid level height estimated by the liquid level height estimation unit 103. For example, the extraction result of the liquid level shape is displayed on a display device or transmitted to an external device via a network.

以下、実施形態における液面形状の抽出方法の詳細を説明する。
図2は、実施形態に係る液面形状の抽出装置100による液面形状の抽出方法を示すフローチャートである。
ステップS1で、入力部101は、カメラ300で撮影した容器200内の自由液面の動画像データを入力する。
Hereinafter, the details of the liquid level shape extraction method in the embodiment will be described.
FIG. 2 is a flowchart showing a method of extracting the liquid level shape by the liquid level shape extraction device 100 according to the embodiment.
In step S1, the input unit 101 inputs the moving image data of the free liquid level in the container 200 taken by the camera 300.

ステップS2で、原データ作成部102は、ステップS1で入力する動画像データを用いて原データを作成する。
図3(a)に、非特許文献2のように動的計画法により曲線を抽出する手法(以下、単に動的計画法と呼ぶ)により作成した原データの例を示す。また、図4(a)に、ある時刻において、原データにおける液面形状401を原画像に重ね合わせて示す。
図4(a)に示すように、画像中のノイズに反応し、液面をなす曲線を誤識別することが原因で、幅方向の広い範囲(図中の左側の領域X)で液面の誤識別が生じている。このように広範囲に渡る誤識別は、空間方向のローパスフィルタ処理等で除去することが困難である。
In step S2, the original data creation unit 102 creates the original data using the moving image data input in step S1.
FIG. 3A shows an example of original data created by a method of extracting a curve by a dynamic programming method (hereinafter, simply referred to as a dynamic programming method) as in Non-Patent Document 2. Further, FIG. 4A shows the liquid level shape 401 in the original data superimposed on the original image at a certain time.
As shown in FIG. 4A, the liquid surface has a wide range in the width direction (region X on the left side in the figure) due to erroneous identification of the curve forming the liquid level in response to noise in the image. Misidentification has occurred. It is difficult to remove such a wide range of misidentifications by low-pass filtering in the spatial direction or the like.

ステップS3で、液面高さ推定部103は、ステップS2で作成した原データを対象として、原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項との重み付き和として評価関数を構築し、この評価関数を最小にする幅方向の各位置、各時刻の液面高さを推定する。
具体的には、図3(a)、図4(a)のような原データを対象として、式(1)の評価関数を最小にする幅方向の各位置j(j=1、2、・・・、M)、各時刻t(t=1、2、・・・、T)の液面高さdj(t)を計算する。yj(t)は、原データにおける液面高さ(動的計画法により求めた液面高さ)である。
In step S3, the liquid level height estimation unit 103 targets the original data created in step S2, and includes a term relating to the difference between the original data and the calculated liquid level height, and the time of the calculated liquid level height. An evaluation function is constructed as a weighted sum of the term related to the difference and the term related to the spatial difference of the calculated value of the liquid level height, and the liquid level height at each position and time in the width direction that minimizes this evaluation function is set. presume.
Specifically, for the original data as shown in FIGS. 3A and 4A, each position j (j = 1, 2, ...) In the width direction that minimizes the evaluation function of the equation (1). ..., M), the liquid level height d j (t) at each time t (t = 1, 2, ..., T) is calculated. y j (t) is the liquid level height in the original data (liquid level obtained by dynamic programming).

Figure 0006759785
Figure 0006759785

式(1)の第1項は液面高さの推定誤差を表わし、第2項は液面の時間方向変動のなめらかさに関する正則化項を表わし、第3項は液面の幅方向変動のなめらかさに関する正則化項を表わす。式(1)は、液面が時間方向及び空間方向にわたりなめらかに変化することを利用して、液面の時間及び空間変動のなめらかさの制約の下で、再構成誤差を最小化するものである。なお、λ1、λ2は推定値変動のなめらかさを制御する調整パラメータである。 The first term of the equation (1) represents the estimation error of the liquid level, the second term represents the regularization term regarding the smoothness of the temporal fluctuation of the liquid level, and the third term represents the width direction fluctuation of the liquid level. Represents a regularization argument for smoothness. Equation (1) minimizes the reconstruction error under the constraint of the smoothness of the temporal and spatial fluctuations of the liquid level by utilizing the fact that the liquid level changes smoothly in the temporal direction and the spatial direction. is there. Note that λ 1 and λ 2 are adjustment parameters that control the smoothness of the estimated value fluctuation.

なお、第2項及び第3項のなめらかさに関する正則化項として、式(1)のような時間方向及び幅方向における液面高さの2階差分の二乗和でなく、式(2)のように、時間方向及び幅方向における液面高さの1階差分の二乗和としてもよい。 As a regularization term for the smoothness of the second and third terms, it is not the sum of squares of the second-order differences of the liquid level in the time direction and the width direction as in the equation (1), but the sum of squares of the difference in the liquid level in the equation (2). As described above, it may be the sum of squares of the first-order differences of the liquid level heights in the time direction and the width direction.

Figure 0006759785
Figure 0006759785

また、動的計画法により求めた液面高さyj(t)は、大きな外れ値を含む場合がある。このため、観測値とのあてはまりの尺度として二乗誤差{yj(t)−dj(t)}2を選んだ場合に、外れ値に推定結果が引きずられ過ぎるおそれがある。これを防ぐため、観測値とのあてはまりの尺度として、二乗誤差でなく誤差絶対値|yj(t)−dj(t)|を選ぶようにしてもよい。この考え方は、ロバスト統計学の分野においてL1損失最小化と呼ばれるものである(非特許文献3を参照)。この場合、式(3)の評価関数を最小にする液面高さdj(t)を計算する。 Further, the liquid level height y j (t) obtained by dynamic programming may include a large outlier. Therefore, when the square error {y j (t) −d j (t)} 2 is selected as a measure of the fit with the observed value, the estimation result may be dragged too much by the outlier. In order to prevent this, the absolute error value | y j (t) −d j (t) | may be selected instead of the squared error as a measure of the fit with the observed value. This idea is called L1 loss minimization in the field of robust statistics (see Non-Patent Document 3). In this case, the liquid level height d j (t) that minimizes the evaluation function of the equation (3) is calculated.

Figure 0006759785
Figure 0006759785

式(3)の数理計画問題は、スラック変数ξj(t)を追加し、式(4)のように書き換えることができる。この変換により、誤差絶対値の項を不等式制約に置き換えることができ、2次計画法を用いて解くことができる。 The mathematical programming problem of equation (3) can be rewritten as in equation (4) by adding the slack variable ξ j (t). By this conversion, the argument of the absolute value of error can be replaced with the inequality constraint, and it can be solved by using the quadratic programming method.

Figure 0006759785
Figure 0006759785

図3(b)に、図3(a)の原データを対象として、式(4)の評価関数を2次計画法を用いて解いて得られた、液面形状の抽出結果(再構成結果)を示す。なお、M=176点、T=12フレーム(0.2sec間)とした。
図3(b)の再構成結果では、図3(a)の原データに含まれる異常値が除去され、時間方向及び空間方向になめらかに変動する液面形状を抽出できている。
なお、この計算にあたり、式(4)の調整パラメータはλ1=λ2=0.5とした。また、計算量低減のため、幅方向に関して10点間隔(約10mm間隔)に間引いて計算し、得られた計算結果を幅方向に補間し最終的な出力としている。
FIG. 3 (b) shows the extraction result (reconstruction result) of the liquid level shape obtained by solving the evaluation function of the equation (4) using the quadratic programming method for the original data of FIG. 3 (a). ) Is shown. In addition, M = 176 points and T = 12 frames (for 0.2 sec).
In the reconstruction result of FIG. 3B, the abnormal value included in the original data of FIG. 3A is removed, and the liquid level shape that smoothly fluctuates in the temporal direction and the spatial direction can be extracted.
In this calculation, the adjustment parameter of Eq. (4) was set to λ 1 = λ 2 = 0.5. Further, in order to reduce the amount of calculation, the calculation is performed by thinning out at 10-point intervals (about 10 mm intervals) in the width direction, and the obtained calculation result is interpolated in the width direction to obtain the final output.

また、図4(b)に、図4(a)と同時刻(12フレーム中の3フレーム目)において、液面形状の再構成結果402を原画像に重ね合わせて示す。図4(b)に示すように、図4(a)で生じていた液面の誤識別が取り除かれ、なめらかな液面形状を抽出できていることがわかる。 Further, FIG. 4B shows the liquid level shape reconstruction result 402 superimposed on the original image at the same time as FIG. 4A (third frame out of 12 frames). As shown in FIG. 4B, it can be seen that the misidentification of the liquid level that occurred in FIG. 4A has been removed and a smooth liquid level shape can be extracted.

ステップS4で、出力部104は、ステップS3で推定した液面高さから得られる液面形状の抽出結果を出力する。出力部104から出力する液面形状の抽出結果としては、例えば図3(b)のようにグラフとして出力してもよいし、図4(b)のように原画像に液面形状の再構成結果402を重ね合わせて出力してもよいし、単に液面高さの値を出力するようにしてもよい。 In step S4, the output unit 104 outputs the extraction result of the liquid level shape obtained from the liquid level height estimated in step S3. As the extraction result of the liquid level shape output from the output unit 104, for example, it may be output as a graph as shown in FIG. 3 (b), or the liquid level shape is reconstructed in the original image as shown in FIG. 4 (b). The result 402 may be superposed and output, or the value of the liquid level may be simply output.

図5に、本発明を適用して液面形状を抽出した結果の例を表わす写真を示す。T=61フレーム(1.0sec間)とし、計算して得られた液面形状の再構成結果を原画像に重ね合わせた結果を6フレーム毎(0.1secピッチ)に10枚表示した例である。図5からもわかるように、時間方向及び空間方向になめらかに変動する液面形状を抽出できている。 FIG. 5 shows a photograph showing an example of the result of extracting the liquid level shape by applying the present invention. In the example where T = 61 frames (for 1.0 sec) and the result of superimposing the reconstruction result of the liquid level shape obtained by calculation on the original image is displayed every 6 frames (0.1 sec pitch). is there. As can be seen from FIG. 5, the liquid level shape that smoothly fluctuates in the temporal direction and the spatial direction can be extracted.

以上述べたように、液面形状の変動は、幅方向、高さ方向及び時間方向の三次元空間においてなめらかな曲面を構成することを見出し、液面の誤識別を含む原データの異常値除去アルゴリズムを構築した。この異常値除去アルゴリズムにより、原データにおいて液面の誤識別が生じていても、その誤識別を取り除いて、時間方向及び空間方向になめらかに変動する液面形状を抽出することができる。
これにより、正確な液体の流動状況、液面挙動を把握することが可能となり、安価なコスト、簡易なセッティングで、連続鋳造機の鋳型を対象とする水モデル実験での水面高さの定量測定ができるようになる。この結果、実機では実施困難な各種実験を可能とし、良好な鋳片品質を可能とする操業因子を明らかにすることができる。
As described above, it was found that the fluctuation of the liquid level shape constitutes a smooth curved surface in the three-dimensional space in the width direction, the height direction and the time direction, and the outliers of the original data including the misidentification of the liquid level are removed. I built an algorithm. With this outlier removal algorithm, even if the liquid level is misidentified in the original data, the misidentification can be removed and the liquid level shape that smoothly fluctuates in the temporal direction and the spatial direction can be extracted.
This makes it possible to accurately grasp the liquid flow status and liquid level behavior, and quantitatively measure the water level in a water model experiment targeting a mold of a continuous casting machine at an inexpensive cost and with simple settings. Will be able to. As a result, it is possible to carry out various experiments that are difficult to carry out with an actual machine, and to clarify the operating factors that enable good slab quality.

以上、本発明を種々の実施形態と共に説明したが、本発明はこれらの実施形態にのみ限定されるものではなく、発明の範囲内で変更等が可能である。例えば図1では、液面形状の抽出装置100において原データを作成するようにしたが、原データは外部で作成するようにして、液面形状の抽出装置100では、作成済みの原データを入力して、液面高さ推定部103で液面高さを推定する構成としてもよい。 Although the present invention has been described above with various embodiments, the present invention is not limited to these embodiments and can be modified within the scope of the invention. For example, in FIG. 1, the original data is created by the liquid level shape extraction device 100, but the original data is created externally, and the created original data is input by the liquid level shape extraction device 100. Then, the liquid level height estimation unit 103 may be configured to estimate the liquid level height.

また、本発明は、連続鋳造機の鋳型の水モデル実験に限定されるものではない。例えば電力会社における原油燃料タンクや原子炉溶融プール等におけるスロッシング現象解明を目的として水モデル実験が行われることがあり、このような場合においても本発明は適用可能である。また、容器200の形状は、用途に合わせたものとすればよい。
さらに、本発明は、水モデル実験に限定されるものではなく、一般に、誤差やノイズを含む動画像から、なめらかな変化を有する自由液面形状を抽出するのに適用可能である。
Further, the present invention is not limited to the water model experiment of the mold of the continuous casting machine. For example, a water model experiment may be conducted for the purpose of clarifying the sloshing phenomenon in a crude oil fuel tank, a nuclear reactor melting pool, or the like in an electric power company, and the present invention is applicable even in such a case. Further, the shape of the container 200 may be adapted to the intended use.
Furthermore, the present invention is not limited to water model experiments, and is generally applicable to extract a free liquid level shape having a smooth change from a moving image including errors and noise.

本発明を適用した液面形状の抽出装置は、例えばCPU、ROM、RAM等を備えたコンピュータ装置により実現される。
また、本発明は、本発明の機能を実現するソフトウェア(プログラム)を、ネットワーク又は各種記憶媒体を介してシステム或いは装置に供給し、そのシステム或いは装置のコンピュータがプログラムを読み出して実行することによっても実現可能である。
The liquid level shape extraction device to which the present invention is applied is realized by, for example, a computer device equipped with a CPU, ROM, RAM, and the like.
The present invention also provides software (programs) that realize the functions of the present invention to a system or device via a network or various storage media, and the computer of the system or device reads and executes the program. It is feasible.

100:液面形状の抽出装置
101:入力部
102:原データ作成部
103:液面高さ推定部
104:出力部
200:容器
201:水
300:カメラ
100: Liquid level shape extraction device 101: Input unit 102: Original data creation unit 103: Liquid level height estimation unit 104: Output unit 200: Container 201: Water 300: Camera

Claims (6)

容器内の自由液面を撮影した、時間的に連続する複数の画像に基づいて液面形状を抽出する液面形状の抽出方法であって、
前記時間的に連続する画像を用いて作成される、各時間において一水平方向の各位置で液面高さが単一に定まる原データを対象として、
前記原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項とを含む評価関数に基づいて、前記一水平方向の各位置、各時刻の液面高さを推定し、
前記原データと液面高さの計算値との差分に関する項は、jを前記一水平方向の各位置(j=1、2、・・・、M)、tを各時刻(t=1、2、・・・、T)とし、前記原データにおける液面高さy j (t)と、液面高さの計算値d j (t)との誤差絶対値|y j (t)−d j (t)|の前記一水平方向の各位置、各時刻に関する和で表わされる、
ことを特徴とする液面形状の抽出方法。
It is a liquid level shape extraction method that extracts the liquid level shape based on a plurality of temporally continuous images of the free liquid level in the container.
For the original data in which the liquid level height is fixed at each position in one horizontal direction at each time, which is created using the temporally continuous images.
An evaluation function that includes a term relating to the difference between the original data and the calculated value of the liquid level, a term relating to the time difference of the calculated value of the liquid level, and a term relating to the spatial difference of the calculated value of the liquid level. Based on this, the liquid level at each position in the one horizontal direction and at each time is estimated .
Regarding the term regarding the difference between the original data and the calculated value of the liquid level, j is each position (j = 1, 2, ..., M) in the one horizontal direction, and t is each time (t = 1, 2, ..., T), and the absolute value of the error between the liquid level height y j (t) in the original data and the calculated liquid level height d j (t) | y j (t) −d It is represented by the sum of each position and each time in the one horizontal direction of j (t) |.
A method for extracting the liquid level shape.
スラック変数を用いて前記誤差絶対値を不等式制約に置き換え、前記評価関数を2次計画法により解くことを特徴とする請求項に記載の液面形状の抽出方法。 The liquid level shape extraction method according to claim 1 , wherein the error absolute value is replaced with an inequality constraint by using a slack variable, and the evaluation function is solved by a quadratic programming method. 前記液面高さの計算値の時間差分に関する項は、時間方向における液面高さの1階差分又は2階差分の二乗和で表わされることを特徴とする請求項1又は2に記載の液面形状の抽出方法。 The liquid according to claim 1 or 2 , wherein the term relating to the time difference of the calculated value of the liquid level is represented by the sum of squares of the first-order difference or the second-order difference of the liquid level in the time direction. Surface shape extraction method. 前記液面高さの計算値の空間差分に関する項は、空間方向における液面高さの1階差分又は2階差分の二乗和で表わされることを特徴とする請求項1乃至のいずれか1項に記載の液面形状の抽出方法。 Any one of claims 1 to 3 , wherein the term relating to the spatial difference of the calculated liquid level height is represented by the sum of squares of the first-order difference or the second-order difference of the liquid level in the spatial direction. The method for extracting the liquid level shape according to the section. 容器内の自由液面を撮影した、時間的に連続する複数の画像に基づいて液面形状を抽出する液面形状の抽出装置であって、
前記時間的に連続する画像を用いて作成される、各時間において一水平方向の各位置で液面高さが単一に定まる原データを対象として、
前記原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項とを含む評価関数に基づいて、前記一水平方向の各位置、各時刻の液面高さを推定する推定手段を備え
前記原データと液面高さの計算値との差分に関する項は、jを前記一水平方向の各位置(j=1、2、・・・、M)、tを各時刻(t=1、2、・・・、T)とし、前記原データにおける液面高さy j (t)と、液面高さの計算値d j (t)との誤差絶対値|y j (t)−d j (t)|の前記一水平方向の各位置、各時刻に関する和で表わされる、
ことを特徴とする液面形状の抽出装置。
It is a liquid level shape extraction device that extracts the liquid level shape based on a plurality of temporally continuous images of the free liquid level in the container.
For the original data in which the liquid level height is fixed at each position in one horizontal direction at each time, which is created using the temporally continuous images.
An evaluation function that includes a term relating to the difference between the original data and the calculated value of the liquid level, a term relating to the time difference of the calculated value of the liquid level, and a term relating to the spatial difference of the calculated value of the liquid level. Based on the above, an estimation means for estimating the liquid level at each position and each time in the one horizontal direction is provided .
Regarding the term regarding the difference between the original data and the calculated value of the liquid level, j is each position (j = 1, 2, ..., M) in the one horizontal direction, and t is each time (t = 1, 2, ..., T), and the absolute value of the error between the liquid level height y j (t) in the original data and the calculated liquid level height d j (t) | y j (t) −d It is represented by the sum of each position and each time in the one horizontal direction of j (t) |.
A liquid level shape extraction device characterized by this.
容器内の自由液面を撮影した、時間的に連続する複数の画像に基づいて液面形状を抽出するためのプログラムであって、
前記時間的に連続する画像を用いて作成される、各時間において一水平方向の各位置で液面高さが単一に定まる原データを対象として、
前記原データと液面高さの計算値との差分に関する項と、液面高さの計算値の時間差分に関する項と、液面高さの計算値の空間差分に関する項とを含む評価関数に基づいて、前記一水平方向の各位置、各時刻の液面高さを推定し、
前記原データと液面高さの計算値との差分に関する項は、jを前記一水平方向の各位置(j=1、2、・・・、M)、tを各時刻(t=1、2、・・・、T)とし、前記原データにおける液面高さy j (t)と、液面高さの計算値d j (t)との誤差絶対値|y j (t)−d j (t)|の前記一水平方向の各位置、各時刻に関する和で表わされる、処理をコンピュータに実行させるためのプログラム。
A program for extracting the liquid level shape based on a plurality of temporally continuous images of the free liquid level in the container.
For the original data in which the liquid level height is fixed at each position in one horizontal direction at each time, which is created using the temporally continuous images.
An evaluation function that includes a term relating to the difference between the original data and the calculated value of the liquid level, a term relating to the time difference of the calculated value of the liquid level, and a term relating to the spatial difference of the calculated value of the liquid level. Based on this, the liquid level at each position in the one horizontal direction and at each time is estimated .
Regarding the term regarding the difference between the original data and the calculated value of the liquid level, j is each position (j = 1, 2, ..., M) in the one horizontal direction, and t is each time (t = 1, 2, ..., T), and the absolute value of the error between the liquid level height y j (t) in the original data and the calculated liquid level height d j (t) | y j (t) −d A program for causing a computer to execute a process represented by the sum of each position and each time in the one horizontal direction of j (t) | .
JP2016137867A 2016-07-12 2016-07-12 Liquid level shape extraction method, equipment and program Active JP6759785B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2016137867A JP6759785B2 (en) 2016-07-12 2016-07-12 Liquid level shape extraction method, equipment and program

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2016137867A JP6759785B2 (en) 2016-07-12 2016-07-12 Liquid level shape extraction method, equipment and program

Publications (2)

Publication Number Publication Date
JP2018009847A JP2018009847A (en) 2018-01-18
JP6759785B2 true JP6759785B2 (en) 2020-09-23

Family

ID=60995329

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2016137867A Active JP6759785B2 (en) 2016-07-12 2016-07-12 Liquid level shape extraction method, equipment and program

Country Status (1)

Country Link
JP (1) JP6759785B2 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6935769B2 (en) * 2018-02-22 2021-09-15 日本製鉄株式会社 State estimation method, molten metal level control method, program, and state estimation device

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3583659B2 (en) * 1999-08-04 2004-11-04 富士通株式会社 Liquid level detection method
US6816436B1 (en) * 2003-05-21 2004-11-09 Siemens Milltronics Process Instruments Inc. Method for echo processing in time-of-flight or level measurement systems
JP4148041B2 (en) * 2003-06-27 2008-09-10 ソニー株式会社 Signal processing apparatus, signal processing method, program, and recording medium
JP2007078483A (en) * 2005-09-13 2007-03-29 Matsushita Electric Ind Co Ltd Liquid level measuring method and liquid quantity measuring method
JP4730954B2 (en) * 2005-09-26 2011-07-20 独立行政法人 日本原子力研究開発機構 Method for obtaining velocity distribution and pressure distribution in a fluid region whose surface shape changes over time
JP4839442B2 (en) * 2006-09-15 2011-12-21 国立大学法人長岡技術科学大学 Image processing system, image processing method, and program
TW200834459A (en) * 2007-02-05 2008-08-16 Huper Lab Co Ltd Video object segmentation method applied for rainy situations
US8718940B2 (en) * 2010-11-30 2014-05-06 Halliburton Energy Services, Inc. Evaluating surface data
US20120281096A1 (en) * 2011-05-02 2012-11-08 Honeywell-Enraf B.V. Storage tank inspection system and method
WO2013088827A1 (en) * 2011-12-16 2013-06-20 旭硝子株式会社 Video image analysis device, video image analysis method, and video image analysis program
FR2985065B1 (en) * 2011-12-21 2014-01-10 Univ Paris Curie OPTICAL FLOAT ESTIMATING METHOD FROM LIGHT ASYNCHRONOUS SENSOR
DE102012215215A1 (en) * 2012-08-27 2014-05-28 Inb Vision Ag Method and device for detecting deviations of a surface of an object

Also Published As

Publication number Publication date
JP2018009847A (en) 2018-01-18

Similar Documents

Publication Publication Date Title
KR100326702B1 (en) Method for calculating road width of the alarm device for lane beyond
TWI774999B (en) A method for monitoring additive manufacturing by an electronic device, a three-dimensional printing device and a related computer-readable medium
JP6737164B2 (en) Liquid surface shape extraction method, device and program
JP2022002111A (en) Manufacturing condition output device, quality management system, and program
US9165341B2 (en) Method for generating super-resolution images having improved image resolution and measuring device
JP6759785B2 (en) Liquid level shape extraction method, equipment and program
JP2019124539A5 (en)
Zappulla et al. Visualization of Steel Continuous Casting Including a New Integral Method for Post‐Processing Temperature Data
JP2019503751A5 (en)
Eslami et al. On how defining and measuring a channel bed elevation impacts key quantities in sediment overloading with supercritical flow
US11865613B2 (en) Structured light part quality monitoring for additive manufacturing and methods of use
JP2015174116A (en) Estimation method for shrinkage crack and memory medium for estimation program of the same
JP6372217B2 (en) Method and apparatus for estimating state of molten metal level in continuous casting mold
Boschetto et al. Powder bed monitoring via digital image analysis in additive manufacturing
KR101905832B1 (en) Method and apparatus for detecting breackout in continuous casting process
JP6818263B2 (en) Fracture surface analysis device and fracture surface analysis method
US11360039B2 (en) Identification of variations or defects in a slab material, made from a continuous casting process
JP2007125589A (en) Apparatus and method for analyzing casting process
Dinda et al. Gas bubble digital data generation by image analysis using reduced-scale water modeling of a slab continuous caster mold
JP5874996B2 (en) Object motion estimation apparatus, object motion estimation method, and program
JP6236687B2 (en) Overhead inspection system using time series filter processing
Lepoittevin et al. Continuous tracking of structures from an image sequence
KR101716208B1 (en) Despiking method of abnormal signal data acquired with using three-dimensional doppler velocimetry in open channel flow and system for the same
JP2007275938A (en) Method for detecting flow-out of slag, method for controlling pouring of molten metal, device for detecting flow-out of slag, device for controlling pouring of molten metal, program, and computer readable recording medium
Michishita et al. Simultaneous Estimation of Optical Flow and Its Boundaries based on the Dynamical System Model

Legal Events

Date Code Title Description
A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20190306

A131 Notification of reasons for refusal

Free format text: JAPANESE INTERMEDIATE CODE: A131

Effective date: 20191119

A02 Decision of refusal

Free format text: JAPANESE INTERMEDIATE CODE: A02

Effective date: 20200317

A521 Written amendment

Free format text: JAPANESE INTERMEDIATE CODE: A523

Effective date: 20200528

A911 Transfer of reconsideration by examiner before appeal (zenchi)

Free format text: JAPANESE INTERMEDIATE CODE: A911

Effective date: 20200604

TRDD Decision of grant or rejection written
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20200804

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20200817

R151 Written notification of patent or utility model registration

Ref document number: 6759785

Country of ref document: JP

Free format text: JAPANESE INTERMEDIATE CODE: R151