TW201035324A - Method for monitoring blast furnace tuyere - Google Patents

Method for monitoring blast furnace tuyere Download PDF

Info

Publication number
TW201035324A
TW201035324A TW98108796A TW98108796A TW201035324A TW 201035324 A TW201035324 A TW 201035324A TW 98108796 A TW98108796 A TW 98108796A TW 98108796 A TW98108796 A TW 98108796A TW 201035324 A TW201035324 A TW 201035324A
Authority
TW
Taiwan
Prior art keywords
image
tuyere
value
target area
axis
Prior art date
Application number
TW98108796A
Other languages
Chinese (zh)
Other versions
TWI394841B (en
Inventor
Qiu-Yi He
Zhi-Zhong Wang
xian-wen Du
Original Assignee
China 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 China Steel Corp filed Critical China Steel Corp
Priority to TW98108796A priority Critical patent/TWI394841B/en
Publication of TW201035324A publication Critical patent/TW201035324A/en
Application granted granted Critical
Publication of TWI394841B publication Critical patent/TWI394841B/en

Links

Abstract

The invention relates to a method for monitoring blast furnace tuyere. The method comprises the following steps (a) capturing a plurality of images for a tuyere of a blast furnace operated at the normal condition; (b) determining a region of interest (ROI) of each image taken in step (a); (c) deciding a baseline based on image data within a region of interest of the image that is performed by step (a); (d) capturing an image of the tuyere to be examined; (e) deciding a region of interest of the image of the tuyere to be examined; (f) calculating image date within the region of interest in the image of the tuyere to be examined; and (g) comparing image date within the region of interest of the image of the tuyere to be examined with the baseline to determine whether or not any abnormal condition occurs. Accordingly, it can automatically set the region of interest of the image of the tuyere and determine whether or not any abnormal condition exists in the tuyere.

Description

201035324 六、發明說明: 【發明所屬之技術領域】 本發明係關於一種監測高爐風口之方法,詳言之,係關 於一種經由風口以監測高爐内是否發生異常之方法。 【先前技術】 參考圖1 ’顯示習知用於高爐之鼓風嘴件組之剖視示竟 圖。該鼓風嘴件組1包括一鼓風管11、一喷煤鎗12、— 規 Ο 〇 視孔13及一側管14。該鼓風管11係插設於一高爐之側壁15 之風口 151上。該喷煤鎗12係連設於該鼓風管丨丨旁用以 將粉煤輸送至該鼓風管U之前端。該窺視孔13係位於該鼓 風官11之後端,用以提供一操作員觀察之管道。該側管Μ 係連通該鼓風管11,用以供一熱風進入該鼓風管u中。 该鼓風管11之前端外附近係形成一風徑區16,當粉煤自 該噴煤搶12被吹出,並隨著該側管14所鼓入之熱風-起在 4風t區16進行燃燒時’該操作員可以從該窺視孔監視 粉煤的噴吹狀況以及該風徑區16的燃燒情況。由於該風徑 區6釔定與否係直接影響該高爐之效率,因此,監測該高 爐風口便是一項重要的課題。 目刖對於^爐之風徑區16之監視主要是以人工巡爐為 為了文。人工直接觀察該窺視孔之不便,有少數之 窥視孔13外加裝一攝影機,且該攝影機再連接至一控制室 榮幕藉此,可供操作員從該控制t m $而,由於 Z系統無自動監測警示功能,控制室之操㈣必須隨時 觀察’否則將難以察覺異况的發生。 138519.doc 201035324 中華民國發明專利第1246537號揭示—種風經區監視方 法及其設備,其乃採用—光料來進行高爐鼓風似徑區 之觀測,用以監視風徑區内的粉煤燃燒情況一般來說, 光度計的取像解析度遠低於攝影系統,因此僅能就燃燒區 域求一平均亮度進行分析,因而其功效並不好。 美國專利第5,223,908號運用光感測器,分析穿透 (Transmitted)與反射(Refiectl〇n)或散射關係, 以計算風口相關參數。美國專利第7,2G9,871號主要是以實 驗方法與維度分析(Dimensi〇nal Analysis),找出相關係 數,以建立數學或經驗模型。美國專利第4,842,253號主要 是用一探針從風口伸入,以觀測燃燒狀況。此種方法雖能 直接觀察燃燒狀況’然對高溫高壓之高爐卻有其實施的危 險性。美國專利第5,481,247號為-高爐風口感測系統,由 直流電源、電位計及光感測器構成,藉由電壓輸出,可以 進行異況#斷。然而’其主要訴求在於電子電路設計與訊 號處理’能使用資訊不多’因此也無法發展較複雜分析工 作。 综觀上述所有方法,其因著重在將每一次觀測到風口狀 況以一簡單的電壓或電流訊號表示,因此在分析資訊不足 的狀況下,將難以進行準確的分析。 因此’有必要提供一創新且富有進步性之監測高爐風口 之方法,以解決上述問題。 【發明内容】 本發明提供一種監測高爐風口之方法,其包括以下步 138519.doc 201035324 驟:(a)擷取複數張該高爐在正常操作下之風口影像;(b) 決定每一張步驟⑷之風口影像之目標區域(Regi〇n of Interest,ROI) ; (c)根據該等步驟(&)之影像之目標區域内之 影像資料決定一基線(Baseline) ; (d)擷取一張待檢測之風 口影像’(e)決定該待檢測之風口影像之目標區域;(〇計 算該待檢測之風口影像之目標區域内之影像資料;及(g)比 對該待檢測之風口影像之目標區域内之影像資料與該基 〇 線,以判斷是否發生異況。藉此,可自動設定風口影像之 目標區域,以及自動辨識風口是否發生異況。 【實施方式】 參考圖2 ’顯示本發明之監測高爐風口之設備之示意 圖。本發明之設備2包括一鼓風嘴件組〗、一攝影機21及一 影像處理裝置22。該鼓風嘴件組1係與圖1之鼓風嘴件組1 相同’其包括該鼓風管11、該喷煤鎗12、該窺視孔13及該 側官14。該攝影機21係位於該窺視孔13外,用以從該窺視 ❹ 孔13擷取該高爐之風口 151之影像。較佳地,該攝影機21 係為一CCD攝影機,其解析度為640x480。該影像處理裝 置22係為一電腦’其内有軟體程式以進行影像處理。 - 參考圖3 ’顯示本發明監測高爐風口之方法之實施例之 々IL程圖。本發明之方法之第一實施例包括以下步驟。步驟 S301係掏取複數張該高爐在正常操作下之風口影像,如圖 4所示。圖4係顯示該攝影機21未歪斜之情況下所擷取之風 口像之示意圖’由於該攝影機21係對準該風口 151,因 此在中央位置會有一近似圓形的亮區(如圖中沒有剖面線 138519.doc 201035324 之區域),其餘部分均為里 “、、色之暗區(如圖_剖面線所在之 冑中上方—個突出物係為該噴煤繪12。可以理解 的是’該噴祕12之前端可能會有喷出之粉煤,而形成此 本貫施例中’係累積1300張以上之風口 影像。 步驟S302係决疋每一張步驟幻^之風口影像之目標區域 ⑽gion of Inte刚,R〇I)。在本發明中’決定出目標區域 之方式包含但不限於以下幾種方式。 第種决疋出该目標區域之方式如下,參考圖# 先,建立該風口影像之凸包(C〇nvex-huU)。接著,利用該 凸包的邊緣迴歸分析,以決U該目標區域。在本實施 例中,由於凸包會略過(Bypass)該喷煤鎗12之輪廓,因此 迴歸後之該目標區域1〇為包圍該亮區之正圓形或橢圓形。201035324 VI. Description of the Invention: [Technical Field] The present invention relates to a method for monitoring a tuyere tuyere, and more particularly to a method for monitoring whether an abnormality occurs in a blast furnace via a tuyere. [Prior Art] A cross-sectional view showing a conventional blaster member group for a blast furnace is shown with reference to Fig. 1'. The blaster member set 1 includes a blast tube 11, a coal injection gun 12, a 〇 〇 sight 13 and a side tube 14. The blast tube 11 is inserted into the tuyere 151 of the side wall 15 of a blast furnace. The coal injection gun 12 is connected to the blast tube to convey the pulverized coal to the front end of the blast tube U. The peephole 13 is located at the rear end of the blaster 11 to provide a conduit for operator observation. The side tube is connected to the blast tube 11 for a hot air to enter the blast tube u. A wind-diameter zone 16 is formed in the vicinity of the front end of the blast tube 11. When the pulverized coal is blown out from the coal-spraying 12, and the hot air blown by the side pipe 14 is started in the 4 wind t zone 16. When burning, the operator can monitor the blowing condition of the pulverized coal and the combustion condition of the wind-diameter area 16 from the peephole. Since the determination of the wind tunnel zone 6 directly affects the efficiency of the blast furnace, monitoring the blast furnace tuyere is an important issue. The monitoring of the wind tunnel area 16 of the furnace is mainly based on the manual inspection furnace. Manually observing the inconvenience of the peephole, a small number of peepholes 13 are attached with a camera, and the camera is connected to a control room, thereby allowing the operator to control the tm$ from the Z system. There is no automatic monitoring warning function, and the operation of the control room (4) must be observed at any time. Otherwise, it will be difficult to detect the occurrence of the abnormal situation. 138519.doc 201035324 The Republic of China invention patent No. 1246537 discloses a method for monitoring a wind-blown area and its equipment, which uses a light material to observe the blast furnace blast-like area for monitoring the pulverized coal in the wind-track area. Combustion In general, the resolution of the photometer is much lower than that of the photographic system, so it can only be analyzed for an average brightness of the combustion area, so its efficacy is not good. U.S. Patent No. 5,223,908 utilizes a light sensor to analyze the relationship between transmitted and reflected (Refiectl〇n) or scattering to calculate tuyere related parameters. U.S. Patent No. 7,2,9,871 is primarily based on Dimensi〇nal Analysis to find correlations to establish mathematical or empirical models. U.S. Patent No. 4,842,253 mainly uses a probe to extend from the tuyere to observe the combustion condition. Although this method can directly observe the combustion condition, the high temperature and high pressure blast furnace has its danger of implementation. U.S. Patent No. 5,481,247 is a blast furnace tuyere sensing system consisting of a DC power source, a potentiometer, and a photo sensor. By voltage output, it can be turned off. However, 'the main appeal is that electronic circuit design and signal processing can't use much information' and therefore it is impossible to develop more complex analysis work. Looking at all the above methods, the emphasis is on a simple voltage or current signal for each observed vent condition, so it will be difficult to perform an accurate analysis under the condition of insufficient analysis information. Therefore, it is necessary to provide an innovative and progressive method of monitoring the blast furnace tuyere to solve the above problems. SUMMARY OF THE INVENTION The present invention provides a method of monitoring a blast furnace tuyere comprising the following steps 138519.doc 201035324: (a) drawing a plurality of tuyere images of the blast furnace under normal operation; (b) determining each step (4) The target area of the image of the tuyere (Regi〇n of Interest, ROI); (c) determine a baseline according to the image data in the target area of the image of the step (&); (d) capture one The image of the tuyere to be detected '(e) determines the target area of the tuyere image to be detected; (calculates the image data in the target area of the tuyere image to be detected; and (g) compares the image of the tuyere to be detected The image data in the target area and the base line are used to determine whether an abnormality has occurred. Thereby, the target area of the tuyere image can be automatically set, and whether the tuyere is automatically recognized or not. [Embodiment] Referring to FIG. 2 The device of the present invention comprises a blaster member set, a camera 21 and an image processing device 22. The blaster member set 1 is coupled to the drum of Fig. 1. The tuyere member set 1 is identical 'which includes the blast tube 11, the coal injection gun 12, the peephole 13 and the side panel 14. The camera 21 is located outside the peephole 13 for viewing from the peephole 13 Preferably, the camera 21 is a CCD camera having a resolution of 640 x 480. The image processing device 22 is a computer having a software program for image processing. Referring to Fig. 3', there is shown a 程IL diagram of an embodiment of the method for monitoring a blast furnace tuyere of the present invention. The first embodiment of the method of the present invention comprises the following steps. Step S301 is to extract a plurality of tuyere images of the blast furnace under normal operation. 4 is a schematic view showing the image of the tuyere taken in the case where the camera 21 is not skewed. Since the camera 21 is aligned with the tuyere 151, there is an approximately circular brightness at the center position. The area (the area without the hatching line 138519.doc 201035324 in the figure), the rest are the dark areas of the ",, color" (as shown in the figure _ the top of the line where the section line is located - a prominent item is the coal injection 12 It is understandable that ' At the front end of the blasting 12, there may be pulverized coal, and in this embodiment, the ventilator image of more than 1300 is accumulated. Step S302 is the target area of the image of the ventilator of each step (10) gion In Inte just, R〇I). In the present invention, the method of determining the target area includes, but is not limited to, the following methods. The first method for determining the target area is as follows. Referring to Figure #, the image of the tuyere is established. a convex hull (C〇nvex-huU). Then, using the edge regression analysis of the convex hull to determine the U target region. In this embodiment, the convex spur will bypass the coal injection gun 12 The contour, so the target area 1 after the regression is a perfect circle or ellipse surrounding the bright area.

第二種決定出該目標區域之方式如下,參考圖6,首 先,計算該風口影像投射於X軸及γ軸之灰階值,以得到 一X方向灰階曲線61及一 Y方向灰階曲線62 ^該又方向灰階 曲線6 1及該Y方向灰階曲線62中間較凸的部份表示累積較 多灰階值高的像素點,往曲線兩側則逐漸減少,直到最兩 邊完全無灰階值高的像素點。 接著’設定一閥值(例如50)。接著,於該X轴定義一第 一位置63及一第二位置64,其中該第一位置63係為該X方 向灰階曲線61左側開始大於該闊值處,該第二位置64係為 該X方向灰階曲線61右側開始小於該閥值處,且於該第一 位置63及該第二位置64分別定義一第一邊65及一第二邊 138519.doc 201035324 66 ’該第一邊65及該第二邊66係平行。同時,於該γ 軸定義一第三位置67及一第四位置68,其中該第三位置67 係為該Υ方向灰階曲線62下側開始大於該閥值處,該第四 位置68係為該γ方向灰階曲線62上侧開始小於該閥值處, 且於該第三位置67及該第四位置68分別定義一第三邊69及 一第四邊7〇,該第三邊09及該第四邊7〇係平行X抽,該第 一邊65、該第二邊66、該第三邊69及該第四邊70形成一矩 形。 以該矩形之中心7 1為圓心,且以該矩形之長及寬之較大 者為直徑,決定出圓形之該目標區域72。 在本發明中’該目標區域的自動搜尋,除可定位後續異 況分析進行之影像區域,也可藉由該目標區域中心變異程 度,觀察該風口 151振動或該攝影機21鎖緊機構之鬆動(圖 2)。 請再參考圖3,步驟S303係根據該等步驟S3〇1之影像之 目t區域内之影像資料決定一基線(Basei jne)。在本實施 例中,該影像資料係為灰階值,且該基線之決定方式如 下。首先,計算每一該等步驟8301之影像之目標區域内之 灰階值之平均值。接著,計算該等平均值之卡特司 (Kurtosis)值。接著,判斷該卡特司(Kurt〇sis)值是否在3 〇 ±0.5之範圍内,若是,則表示該等平均值之分佈接近常態 分佈(Normal Distribution),此時再將該等平均值取一平均 值及標準差,且定義該等平均值之平均值為該基線。若該 卡特司(Kurtosis)值不在3.0±0.5之範圍内,則返回該步驟 138519.doc 201035324 S301 ’重新再擷取風口影, 京像直到該卡特司(Kurtosis)值 落在3.0土0.5之範圍内為止。 步驟S304係擷取一張待檢、、丨 取侍檢測之風口影像。如同上述,利 用該攝影機21對準該風口】s丨& 凤口 151擷取—張待檢測之風口影 像。步驟S305係決定該待檢带丨 付檢測之風口景》像之目標區域,其 方式與上述決定出目檁區诚 L域之方式相同。步驟S3 06係計算 該待檢測之風口影像之日辨π a + 目標區域内之影像資料(例如灰階 之平均值)。The second way to determine the target area is as follows. Referring to FIG. 6, first, the gray scale value of the tuyere image projected on the X-axis and the γ-axis is calculated to obtain an X-direction gray scale curve 61 and a Y-direction gray scale curve. 62 ^ The convex portion of the gray-scale curve 6 1 and the Y-direction gray-scale curve 62 indicates that more pixels with higher gray-scale values are accumulated, and gradually decrease toward both sides of the curve until the two sides are completely gray-free. A pixel with a high order value. Then 'set a threshold (for example 50). Next, a first position 63 and a second position 64 are defined in the X-axis, wherein the first position 63 is that the left side of the X-direction gray scale curve 61 is greater than the threshold value, and the second position 64 is the The right side of the X-direction gray-scale curve 61 starts to be smaller than the threshold, and the first side 63 and the second position 64 respectively define a first side 65 and a second side 138519.doc 201035324 66 'the first side 65 And the second side 66 is parallel. At the same time, a third position 67 and a fourth position 68 are defined in the γ-axis, wherein the third position 67 is that the lower side of the 灰-direction gray-scale curve 62 starts to be greater than the threshold, and the fourth position 68 is The upper side of the γ-direction gray-scale curve 62 starts to be smaller than the threshold, and the third side 67 and the fourth position 68 respectively define a third side 69 and a fourth side 7〇, the third side 09 and The fourth side 7 is parallel X drawn, and the first side 65, the second side 66, the third side 69 and the fourth side 70 form a rectangle. The center of the rectangle 7 1 is the center of the circle, and the larger of the length and the width of the rectangle is the diameter, and the target area 72 of the circle is determined. In the present invention, the automatic search of the target area, in addition to the image area in which the subsequent out-of-state analysis can be performed, can also observe the vibration of the tuyere 151 or the loosening of the locking mechanism of the camera 21 by the degree of variation of the center of the target area ( figure 2). Referring to FIG. 3 again, step S303 determines a baseline according to the image data in the region t of the image of the step S3〇1. In this embodiment, the image data is a grayscale value, and the baseline is determined as follows. First, an average of the grayscale values in the target region of the image of each of the steps 8301 is calculated. Next, the Kurtosis values of the average values are calculated. Next, it is determined whether the value of the Kurt〇sis is within a range of 3 〇±0.5, and if so, it indicates that the distribution of the average values is close to a normal distribution, and then the average value is taken one at a time. The mean and standard deviation, and the average of the averages is defined as the baseline. If the Kurtosis value is not within the range of 3.0 ± 0.5, return to step 138519.doc 201035324 S301 'Re-take the wind shadow, Jing Jing until the Kurtosis value falls within the range of 3.0 ± 0.5 So far. In step S304, an image of the tuyere to be inspected and taken out is taken. As described above, the camera 21 is used to align the tuyere s丨& Fengkou 151 to capture the image of the tuyere to be detected. Step S305 is to determine the target area of the image of the tuyere of the to-be-checked zone, in the same manner as the above-mentioned manner of determining the area of the area. Step S3 06 calculates the image data of the target image to be detected π a + in the target area (for example, the average value of the gray scale).

步驟㈣係、比對該待檢測之風口影像之目標區域内之影 像資料與該基線,以判斷是否發生異況。如果待檢測之灰 階平均值過高,那有可能是喷煤不足所造成;反之,如果 待檢測之灰(¾平均值偏低,則可能是喷煤過量、大塊落料 或倒灌浮㈣狀況。當大Μ料之異況發生時,會於風口 影像上方區域出現大面積之暗區;當倒灌浮渣之異泥發生 時會於風口影像下方區域出現大面積之暗區;其餘異況 則判斷為喷煤過量。 以下兹舉幾個實施例說明倒灌浮渔之判斷方式,可以理 解的是相同之方式也可以應用於大塊落料之判斷方式,因 為其二者僅是上下位置不同而已。 ,第-種倒灌浮逢之判斷方式如下。首先,利用一ccd攝 影機擷取一張待檢測之風口影像,如圖7所示。接著,決 疋該待檢測之風口影像之亮區之下方輪廓曲線73,其中該 輪廓曲線73係由複數個點組成,且每一點具有一χ座標值 及Υ座標值。接著,計算該等點之γ座標值之卡特司 138519.doc 201035324 (Kurtosis)值。最後,判斯镎上 J所D亥卡特司(Kurtosis)值是否超過 一闕值(例如3·5)’若否’則判斷為不發生倒灌浮逢;若 是,則判斷為發生倒灌浮漬。這是由於當發生倒灌浮法 時’該下方輪廓曲線73會變的十分不規則。 第二種倒灌浮潰之判斷方式如下。首先,利用一咖攝 影機掏取-張待檢測之風口影像,如圖8所示。接著,設 定一水平基準線74。接著,呔中—4 ^ 〇 者决疋该待檢測之風口影像之亮 Ο ❹Step (4) is to compare the image data in the target area of the image of the tuyere to be detected with the baseline to determine whether an abnormality has occurred. If the average gray level to be detected is too high, it may be caused by insufficient coal injection; conversely, if the ash to be detected (the average value of 3⁄4 is low, it may be excessive coal injection, large blanking or inverted floating (4) Situation: When the situation of the big dip occurs, a large area of dark area will appear in the upper part of the image of the tuyere; when the mud of the residing slag occurs, a large area of dark area will appear in the area below the tuyere image; Then, it is judged that the coal injection is excessive. The following describes the judgment method of the reverse floating fishing, and it can be understood that the same method can also be applied to the judgment method of the large blanking because the two are only different in the upper and lower positions. However, the first type of back-floating method is judged as follows. First, a ccd camera is used to capture an image of the tuyere to be detected, as shown in Fig. 7. Then, the bright area of the image of the tuyere to be detected is determined. The lower contour curve 73, wherein the contour curve 73 is composed of a plurality of points, and each point has a coordinate value and a Υ coordinate value. Then, the Cartesian value of the gamma coordinate value of the points is calculated 138519.doc 201035324 (K Urtosis). Finally, whether the Kurtosis value of J is higher than a ( value (for example, 3·5) 'If no', it is judged that no back-filling occurs; if so, it is judged to occur Inverted floating stains. This is because when the floating float method occurs, the lower contour curve 73 will become very irregular. The second type of backflow float is judged as follows. First, use a coffee camera to capture the sample to be detected. The tuyere image is as shown in Fig. 8. Next, a horizontal reference line 74 is set. Then, 呔中-4^ 疋 determines the brightness of the image of the tuyere to be detected Ο

Sit:廓曲線73。接著’求出該下方輪廓曲線73與該 水:基準線74所圍出之亮區面積。接著,利用凸包的邊緣 迴歸分析’以決定出該待檢測之風口影像之亮區之下方凸 凸㈣,求出該下方凸包邊75與該水平基準線⑽ =之1包面積。最後,計算該亮區面積及該凸包面積之 ” ’以差異大於—預設值時即判斷為^生倒灌浮法。 倒灌浮渣之判斷方式如下。首先’利用—c I機掏取-張待檢測之風口影像,如圖9所示。接著,決 =待檢測之風σ影像之亮區之下方輪廓曲線η。接著:、 :、:亥待檢測之風口影像之目標區域之下方外圍曲線%。 “上迷’該下方外圍曲線76之決定方式可以有 二 種。 — …第一種該下方外圍曲線76之決定方式如下,首先建立 ^檢測之風口影像之凸包(c〇nvexhuU)。接者利用該 二:緣迴歸分析’以決定出該目標區域。最後,選取 抓區域之下方外圍曲線76。 第二種該下方外圍曲線76之決定方式如下,本方式_ 138519.doc 201035324 用3亥目;jif;區域,而決定該目標區域之方式如同上述。首 先,計算該風口影像投射於X轴及Y軸之灰階值,以得到 一X方向灰階曲線及一γ方向灰階曲線。接著,設定一閥 值。接著,於該X轴定義一第一位置及一第二位置,其中 3亥第一位置係為該X方向灰階曲線左側開始大於該閥值 處,該第一位置係為該X方向灰階曲線右侧開始小於該閥 值處,且於該第一位置及該第二位置分別定義一第一邊及 ^ 一第二邊’該第一邊及該第二邊係平行Y轴。 同時,於該Y軸定義一第三位置及一第四位置,其中該 第二位置係為該γ方向灰階曲線下側開始大於該閥值處, 該第四位置係為該γ方向灰階曲線上側開始小於該閥值 處’且於該第三位置及該第四位置分別定義一第三邊及一 第四邊,該第三邊及該第四邊係平行乂軸,該第一邊、該 第二邊、該第三邊及該第四邊形成一矩形。 最後,以該矩形之中心為圓心,且以該矩形之長及寬之 〇 較大者為直徑,決定出圓形之該目標區域,且選取該目標 區域之下方外圍曲線76。 最後’計算該下方輪廓曲線73之平均值,且計算該平均 值及該下方外圍曲線7 6最低點間之距離d,當該距離d大於 一預設值時即判斷為發生倒灌浮渣。 本發明之優點為,可自動設定風口影像之目標區域,以 及自動辨識風口是否發生異況,並於必要時並發出警示訊 息。本發明除可建立風口全面自動監視與預警系統外,也 有助於兩爐粉煤喷煤(pulverised C〇ai injecti〇n, PCI)穩定 138519.doc 201035324 的提昇。 上述實施例僅為說明本發明之原理及其功效,並非限制 本發明。因此習於此技術之人士對上述實施例進行修改及 變化仍不脫本發明之精神。本發明之權利範圍應如後述之 申請專利範圍所列。 【圖式簡單說明】 圖1顯示習知用於高爐之鼓風嘴件組之剖視示意圊; ❹ Ο 圖2顯示本發明之監測高爐風口之設備之示意圖; 圖3顯不本發明監測高爐風口之方法之實施例之流程 raj · 圖, 圖4顯示攝影機未歪斜之情況下所擷取之風口影像之示 意圖; 圖5顯示本發明中建立風口影像之凸包(c〇nvex_hu⑴之 不意圖, 圖6顯示本發明中風口影像投射於又轴及γ轴之灰階值之 不意圖, 圖7顯示本發明中第一種倒灌浮渣之判斷方式之示意 方式之示意 rgi , 圖, 圖8顯示本發明中第 圖;及 二種倒灌浮渣之判斷 圖 圖。 9顯示本發明中第三種倒灌浮渣之判斷 方式之示意 【主要元件符號說明】 習知鼓風嘴件組 I38519.doc 201035324 Ο 〇 2 本發明之監測高爐風口之設備 10 目標區域 11 鼓風管 12 噴煤鎗 13 窺視孔 14 側管 15 高爐之側壁 16 風徑區 21 攝影機 22 影像處理裝置 61 X方向灰階曲線 62 Υ方向灰階曲線 63 第一位置 64 第二位置 65 第一邊 66 第二邊 67 第三位置 68 第四位置 69 第三邊 70 第四邊 71 矩形之中心 72 目標區域 73 風口影像之亮區之下方輪廓曲線 74 水平基準線 138519.doc 12- 201035324 75 風口影像之免區之下方凸包邊 76 風口影像之目標區域之下方外圍曲線 151 風口Sit: Profile 73. Next, the lower contour curve 73 and the bright area of the water: reference line 74 are determined. Next, the edge convex regression analysis of the convex hull is used to determine the lower convex portion (four) of the bright region of the tuyere image to be detected, and the lower convex rim 75 and the horizontal reference line (10) = 1 packet area are obtained. Finally, the calculation of the area of the bright area and the area of the convex hull is judged to be the inverted floating float method when the difference is greater than the preset value. The method for determining the scum of the backing is as follows: First, use the -c I machine to capture - The image of the tuyere to be detected is as shown in Fig. 9. Then, the contour curve η below the bright area of the wind image to be detected is determined. Then: , ::: the lower periphery of the target area of the image of the tuyere to be detected Curve %. There are two ways to determine the lower peripheral curve 76. - ... The first type of lower peripheral curve 76 is determined as follows. First, a convex hull (c〇nvexhuU) of the detected air vent image is established. The receiver uses the two: edge regression analysis to determine the target area. Finally, select the peripheral curve 76 below the capture area. The second method of determining the lower peripheral curve 76 is as follows. In this manner, the method of determining the target area is as described above by using the method of 3 hiring; jif; First, the grayscale value of the tuyere image projected on the X-axis and the Y-axis is calculated to obtain an X-direction gray-scale curve and a γ-direction gray-scale curve. Next, set a threshold. Next, a first position and a second position are defined in the X-axis, wherein the first position of the 3H is greater than the threshold value at the left side of the X-direction gray-scale curve, and the first position is the X-direction gray level The right side of the curve begins to be smaller than the threshold value, and a first side and a second side are respectively defined in the first position and the second position. The first side and the second side are parallel to the Y axis. At the same time, a third position and a fourth position are defined in the Y axis, wherein the second position is that the lower side of the gamma gray scale curve starts to be greater than the threshold, and the fourth position is the gamma gray scale. The upper side of the curve begins to be smaller than the threshold value, and a third side and a fourth side are respectively defined in the third position and the fourth position, and the third side and the fourth side are parallel to the first axis. The second side, the third side and the fourth side form a rectangle. Finally, the center of the rectangle is centered, and the larger of the length and width of the rectangle is the diameter, the target area of the circle is determined, and the lower peripheral curve 76 of the target area is selected. Finally, the average value of the lower contour curve 73 is calculated, and the distance d between the average value and the lowest point of the lower peripheral curve 71 is calculated. When the distance d is greater than a predetermined value, it is determined that the scum is poured. The invention has the advantages that the target area of the tuyere image can be automatically set, and whether the tuyere is automatically recognized or not, and a warning message is issued when necessary. The invention not only can establish a comprehensive automatic monitoring and early warning system for the tuyere, but also contributes to the improvement of the pulverised C〇ai injecti〇n (PCI) stability 138519.doc 201035324. The above embodiments are merely illustrative of the principles and effects of the invention and are not intended to limit the invention. Therefore, those skilled in the art can make modifications and changes to the above embodiments without departing from the spirit of the invention. The scope of the invention should be as set forth in the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic cross-sectional view showing a conventional blaster assembly for a blast furnace; FIG. 2 is a schematic view showing the apparatus for monitoring a blast furnace tuyere of the present invention; FIG. 3 is a view showing the blast furnace of the present invention. The flow of the embodiment of the method of the tuyere raj · Fig. 4 shows a schematic view of the image of the tuyere taken when the camera is not skewed; FIG. 5 shows the convexity of the image of the tuyere in the present invention (c〇nvex_hu(1) is not intended, 6 is a schematic view showing the gray scale value of the tuyere image projected on the axis and the γ axis of the present invention, and FIG. 7 is a schematic diagram showing the schematic manner of the first method for judging the scum of the present invention. FIG. The figure of the present invention; and the judgment diagram of the two kinds of backflow scum. 9 shows the mode of judging the third type of back scum in the present invention. [Main component symbol description] Conventional blast nozzle group I38519.doc 201035324 Ο 〇2 Equipment for monitoring blast furnace tuyere of the present invention 10 Target area 11 Blast tube 12 Coal injection gun 13 Peephole 14 Side tube 15 Blast furnace side wall 16 Wind path area 21 Camera 22 Image processing unit 61 X direction gray scale curve 62 Υ direction gray scale curve 63 First position 64 Second position 65 First side 66 Second side 67 Third position 68 Fourth position 69 Third side 70 Fourth side 71 Center of rectangle 72 Target Area 73 Below the contour of the bright image of the tuyere image 74 Horizontal reference line 138519.doc 12- 201035324 75 Below the convex edge of the tuyere image free area 76 Below the peripheral area of the target area of the tuyere image 151 Tuyere

138519.doc -13-138519.doc -13-

Claims (1)

201035324 七、申請專利範圓: 樘監测高爐風 --- ·ιυ 々s 厂 少-辱取 · ⑷操取複數張該高爐在正常操作下Μ口影像; W歧每-張步驟⑷之風σ影像之目標區域(Regi〇n 〇f Interest, R〇i); (c) 根據该荨步驟(a)之影像之h # π u — 、X灸彳豕l目橾q域内之影像資料決 定一基線(Baseline); (d) 擷取一張待檢測之風口影像;201035324 VII. Application for patents: 樘 高 樘 樘 - - - - 樘 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 厂 ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( Target area of σ image (Regi〇n 〇f Interest, R〇i); (c) According to the image data of h # π u — and X moxibustion in the image of step (a) a baseline (Baseline); (d) capture an image of the tuyere to be tested; (e) 決定該待檢測之風口影像之目標區域; (f) 計算該待檢測之風口影像之目標區域内之影像資 料;及 (g)比對該待檢測之風口影像之目標區域内之影像資料 與該基線’以判斷是否發生異況。 2. 如請求項1之方法,其中該步驟(&)及該步驟(d)係利用一 攝影機擷取風口影像。 3. 如請求項2之方法,其中該攝影機係為一 CCD攝影機。 4. 如凊求項1之方法,其中該步驟(b)包括: (bl)建立§亥風口影像之凸包(c〇nvex_hun);及 (b2)利用該凸包的邊緣迴歸分析,以決定出該目標區 域。 5. 如請求項1之方法,其中該步驟(b)包括: (bl)計算該風口影像投射於X軸及γ軸之灰階值,以得 到一X方向灰階曲線及一γ方向灰階曲線; (b2)設定一閥值; I38519.doc 201035324 (b3)於該X轴定義一第一位置及一第二位置,其令該第 一位置係為該X方向灰階曲線左側開始大於該閥值 處’該第二位置係為該X方向灰階曲線右側開始小 於該閥值處’且於該第一位置及該第二位置分別定 義一第一邊及一第二邊,該第一邊及該第二邊係平 行Y軸; (b4)於該Y軸定義一第三位置及一第四位置其令該第 〇 二位置係為該Y方向灰階曲線下側開始大於該閥值 處’該第四位置係為該γ方向灰階曲線上側開始小 於該閥值處’且於該第三位置及該苐四位置分別定 義一第二邊及一第四邊,該第三邊及該第四邊係平 盯X轴,該第一邊、該第二邊、該第三邊及該第四 邊形成一矩形;及 (b5)以》亥矩形之中心為圓心,且以該矩形之長及寬之較 大者為直徑,決定出圓形之該目標區域。 〇 6.如請求们之方法’其中該步驟⑷令之影像資料係為灰 階值。 7.如凊求項1之方法,其中該步驟(〇包括: (cl)。十算每一該等步驟⑷之影像之目標區域内之灰階值 之平均值; (c2)計算該等平均值之卡特司(Kurt〇sis)值;及 (c3)判斷該卡特司(Kurt〇sis)值是否在3 〇±〇 5之範圍 内,若是,則定義所有該等平均值之平均值為該基 線;若否,則返回該步驟(3)。 1385I9.doc 201035324 8. -種監測高爐風口之方法,包括以下步驟: (a) 擷取一張待檢測之風口影像; (b) 決疋該待檢測 M办彳冢之亮區之下方輪廓曲線, J該輪廓曲線係由複數個點組成,且每一點具有 X座標值及一 γ座標值; ⑷計算該等點之γ座標值之卡特司(Kun〇sis)值;及 ⑷判斷該卡特司(Kurtosis)值是否超過一間值,若否, Ο Ο 則判斷為不發生倒灌耗;若是,則判斷為發 灌浮渣。 9’如请求項8之方法,其中該步驟⑷係利用一攝影機擷取 風口影像。 月求項9之方法,其中該攝影機係為一 CCD攝影機。 11. -種監測高爐風口之方法,包括以下步驟: (a) 擷取一張待檢測之風口影像; (b) 設定一水平基準線; ㈠决疋4待檢測之風σ影像之亮區之下方輪廓曲線; (d)求出該下方輪廓曲線與該水平基準線所圍出之亮區 面積; ⑷利用凸包的邊緣迴歸分析,以決定出該待檢測之風 口影像之亮區之下方凸包邊; (f)求出該下方凸包邊與該水平基準線所圍出之凸包面 積; (g)计算該亮區面積及該凸包面積之差異,當該差異大 於一預設值時即判斷為發生倒灌浮渣。 138519.doc 201035324 12, 如請求項u之方法, π “ 、中b驟(a)係利用一攝影機操取 風口影像。 13. 如請求項12之方法’其中該攝影機係為一 CCD攝影機。 種監測高爐風口^心㈣q㈣:免機 (a)擷取一張待檢測之風口影像; (啸該待檢測之風口影像之亮區之下方輪廓曲線; (c) 決定該待檢測之風口 办彳豕(目知區域之下方外圍曲 Ο 線;及 (d) 計算該下方輪廓曲 琛及该下方外圍曲線最低點間之 距離,當該距離大於-預設值時即判斷為發生倒灌 浮渣。 β长項14之方法’其中該步驟⑷係利用—攝影機掏取 風口影像。 %如請求項15之方法,其中該攝影機係為―咖攝影機。 17,如請求項丨4之方法,其中該步驟(c)包括: 〇 (Cl)建立該待檢測之風口影像之凸包(Convex-him);及 (c2)利用该凸包的邊緣迴歸分析’以決定出該目標區 域,及 (c3)選取該目標區域之下方外圍曲線。 1 8.如清求項丨4之方法,其中該步驟(c)包括: (cl)計算該風口影像投射於又軸及γ軸之灰階值,以得 到一X方向灰階曲線及一 γ方向灰階曲線; (c2)設定一閥值; (c3)於該X軸定義一第一位置及一第二位置,其中該第 I38519.doc 201035324 一位置係為該X方向灰階曲線左側開始大於該閥值 處’該第二位置係為該X方向灰階曲線右側開始小 於該閥值處,且於該第一位置及該第二位置分別定 義一第一邊及一第二邊,該第一邊及該第二邊係爭 行Y軸; (c4)於該Y軸定義一第三位置及一第四位置,其中該笫 三位置係為該Y方向灰階曲線下側開始大於該閥值 處,該第四位置係為該Y方向灰階曲線上側開始小 〇 ^ 於該閥值處,且於該第三位置及該第四位置分別定 義一第三邊及一第四邊,該第三邊及該第四邊係平 行X軸’該第一邊、該第二邊、該第三邊及該第四 邊形成一矩形;及 (c5)以該矩形之中心為圓心,且以該矩形之長及寬之較 大者為直徑,決定出圓形之該目標區域,且選取該 目標區域之該下方外圍曲線。 〇 138519.doc(e) determining a target area of the image of the tuyere to be detected; (f) calculating image data in the target area of the image of the tuyere to be detected; and (g) comparing the image in the target area of the image of the tuyere to be detected The data is linked to the baseline to determine if an abnormality has occurred. 2. The method of claim 1, wherein the step (&) and the step (d) utilize a camera to capture the tuyere image. 3. The method of claim 2, wherein the camera is a CCD camera. 4. The method of claim 1, wherein the step (b) comprises: (bl) establishing a convex hull of the 亥海风口 image (c〇nvex_hun); and (b2) using the edge regression analysis of the convex hull to determine Out of the target area. 5. The method of claim 1, wherein the step (b) comprises: (bl) calculating a gray scale value of the tuyere image projected on the X-axis and the γ-axis to obtain an X-direction gray scale curve and a γ-direction gray scale (b2) setting a threshold; I38519.doc 201035324 (b3) defining a first position and a second position on the X axis, such that the first position is greater than the left side of the X direction gray scale curve At the threshold, the second position is that the right side of the X-direction gray-scale curve starts to be smaller than the threshold value, and a first side and a second side are respectively defined in the first position and the second position, the first position The side and the second side are parallel to the Y axis; (b4) defining a third position and a fourth position on the Y axis, wherein the second position is such that the lower side of the Y direction gray scale curve starts to be greater than the threshold Wherein the fourth position is that the upper side of the gamma-direction gray-scale curve starts to be smaller than the threshold value, and a second side and a fourth side are respectively defined at the third position and the fourth four-position, the third side and The fourth side is flat on the X axis, and the first side, the second side, the third side, and the fourth side form a rectangle And (b5) to the center "Hai rectangles as the center, and with the length and width of the rectangle is greater than the diameter, the decision of a circular target area. 〇 6. As requested by the method of the requester, the image data of the step (4) is grayscale value. 7. The method of claim 1, wherein the step (〇) comprises: (cl). calculating the average of the grayscale values in the target region of the image of each of the steps (4); (c2) calculating the average a value of the Kurt〇sis value; and (c3) determining whether the value of the Kurt〇sis is within a range of 3 〇 ± 〇 5, and if so, defining an average of all such average values as Baseline; if not, return to step (3). 1385I9.doc 201035324 8. - Method for monitoring blast furnace tuyere, comprising the following steps: (a) taking an image of the tuyere to be tested; (b) The lower contour curve of the bright area to be detected, J is composed of a plurality of points, and each point has an X coordinate value and a γ coordinate value; (4) Calculating the gamma coordinate value of the points (Kun〇sis) value; and (4) determine whether the Kurtosis value exceeds a value, and if not, Ο Ο judges that no backflow consumption occurs; if yes, it is judged to be scum. 9' The method of item 8, wherein the step (4) utilizes a camera to capture an image of the tuyere. The method of claim 9, wherein the camera is a CCD camera. 11. A method for monitoring a blast furnace tuyere, comprising the steps of: (a) capturing an image of a tuyere to be inspected; (b) setting a horizontal reference line; Determine the lower contour curve of the bright region of the wind σ image to be detected; (d) find the lower contour curve and the area of the bright area enclosed by the horizontal reference line; (4) use the edge regression analysis of the convex hull to determine (b) determining the area of the convex hull surrounded by the lower convex hem and the horizontal reference line; (g) calculating the area of the bright area and the convex The difference in the area of the package, when the difference is greater than a predetermined value, it is judged that the dumping scum occurs. 138519.doc 201035324 12, as in the method of claim u, π ", middle b (a) is operated by a camera 13. The method of claim 12, wherein the camera is a CCD camera. The monitoring blast furnace tuyere (4) q (4): free (a) captures an image of the tuyere to be detected; Below the outline of the bright area of the tuyere image (c) determining the tuyere to be inspected (the peripheral curved line below the target area; and (d) calculating the distance between the lower contour curve and the lowest point of the lower peripheral curve, when the distance is greater than - The default value is judged to be the occurrence of the dumping scum. The method of the β long term 14 wherein the step (4) utilizes the camera to capture the tuyere image. %. The method of claim 15, wherein the camera is a "Caf camera." 17. The method of claim 4, wherein the step (c) comprises: 〇 (Cl) establishing a convex envelope (Convex-him) of the image of the tuyere to be detected; and (c2) performing edge regression analysis using the convex hull 'To determine the target area, and (c3) to select the lower peripheral curve of the target area. 1 8. The method of claim 4, wherein the step (c) comprises: (cl) calculating a gray scale value of the tuyere image projected on the axis and the γ axis to obtain an X-direction gray-scale curve and a γ a direction gray scale curve; (c2) setting a threshold value; (c3) defining a first position and a second position on the X axis, wherein the I38519.doc 201035324 position is the left side of the X direction gray scale curve The second position is greater than the threshold value at the right side of the X-direction gray-scale curve, and a first side and a second side are respectively defined in the first position and the second position. The first side and the second side compete for the Y axis; (c4) defines a third position and a fourth position on the Y axis, wherein the third position is that the lower side of the Y direction gray scale curve starts to be larger than the At the threshold, the fourth position is that the upper side of the Y-direction gray-scale curve starts to be smaller than the threshold, and a third side and a fourth side are respectively defined in the third position and the fourth position. The third side and the fourth side are parallel to the X axis 'the first side, the second side, the third side, and the fourth side A rectangle; and (C5) to the center of the rectangle of a circle, and with the length and width of the rectangle is greater than the diameter, the circular decision of the target area, and select the target area of the peripheral curved downward. 〇 138519.doc
TW98108796A 2009-03-18 2009-03-18 Methods for monitoring blast furnace tuyere TWI394841B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW98108796A TWI394841B (en) 2009-03-18 2009-03-18 Methods for monitoring blast furnace tuyere

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW98108796A TWI394841B (en) 2009-03-18 2009-03-18 Methods for monitoring blast furnace tuyere

Publications (2)

Publication Number Publication Date
TW201035324A true TW201035324A (en) 2010-10-01
TWI394841B TWI394841B (en) 2013-05-01

Family

ID=44855768

Family Applications (1)

Application Number Title Priority Date Filing Date
TW98108796A TWI394841B (en) 2009-03-18 2009-03-18 Methods for monitoring blast furnace tuyere

Country Status (1)

Country Link
TW (1) TWI394841B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111860429A (en) * 2020-07-30 2020-10-30 科大讯飞股份有限公司 Blast furnace tuyere abnormality detection method, device, electronic apparatus, and storage medium

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN2913380Y (en) * 2006-06-22 2007-06-20 重庆大学 On-line monitoring device for blast furnace tuyere operating mode
TWI316863B (en) * 2007-03-15 2009-11-11 Nat Univ Chin Yi Technology Sphere identification system

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111860429A (en) * 2020-07-30 2020-10-30 科大讯飞股份有限公司 Blast furnace tuyere abnormality detection method, device, electronic apparatus, and storage medium
CN111860429B (en) * 2020-07-30 2024-02-13 科大讯飞股份有限公司 Blast furnace tuyere abnormality detection method, device, electronic equipment and storage medium

Also Published As

Publication number Publication date
TWI394841B (en) 2013-05-01

Similar Documents

Publication Publication Date Title
JP5672480B2 (en) Apparatus and method for determining shape of terminal end of bead
TWI512110B (en) Abnormal detection methods and blast furnace operation methods
CN103743761B (en) A kind of eyeglass watermark defect image detection device
JP6457137B1 (en) Garbage mixing degree evaluation system
KR20130049826A (en) Gob inspection system for glass product
TWI512284B (en) Bubble inspection system for glass
JP5263051B2 (en) Laser irradiation point detection device and seam position detection device in laser welding machine
KR100939424B1 (en) Method for Detecting Defect in Glass Panel
TW201035324A (en) Method for monitoring blast furnace tuyere
JP5732605B2 (en) Appearance inspection device
TW201617605A (en) Defect inspection method and apparatus thereof
JP2011075325A (en) Surface inspection device
JP4821524B2 (en) Projection height measurement method, projection height measurement device, program
JP2004361085A (en) Visual examination device
JP2010071980A (en) Fin extraction inspection method of heat exchanger
GB2390675A (en) Flame characteristic monitor using digitising image camera
JP2015210151A (en) Method and device for detecting surface defect
TW201239100A (en) Method for monitoring furace through tuyeres
JPH07260696A (en) Method for detecting slag run-off when unloading steel from converter
JP2006292500A (en) Surface inspection method and surface inspection device
JP3454162B2 (en) Inspection system for flat plate with holes
JP2004286708A (en) Defect detection apparatus, method, and program
JP7017973B2 (en) Haesagari determination device and method
JP5376355B2 (en) Winding defect inspection machine
JP3846840B2 (en) Inspection method and apparatus for electronic display

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees