TWI699706B - Method for detecting liquid level turbulence of surface image - Google Patents

Abstract

A method for detecting a liquid level turbulence of a surface image includes steps of: (a) providing a mold level; (b) providing a first image frame, a second image frame and a third image frame of the mold level in time intervals; (c) defining a second region around the first region of the image frames, wherein the second region is larger than the first region; (d) searching for a region similar to the image of the first region of the second image frame in the second region of the first image frame to calculate a first similarity; and (e) converting the first similarity to a degree of difference, to find a boiling probability, thereby determining a degree of boiling in the second region of the first image frame .

Description

Surface image liquid level disturbance detection method

The invention relates to a detection method for surface image liquid level disturbance, in particular to a surface image liquid level disturbance detection method for continuous casting mold liquid level in a steelmaking plant.

The occurrence of hot rolling defects (steel-making quality defects) in steel mills has remained high in recent years, which is a major problem in steel product quality. It is known that the sudden boiling of the mold liquid level is a major cause of slag entrapment. The current mold liquid level detection methods mostly use electromagnetic sensing to detect the state of the entire mold liquid level with a single-point position disturbance. Moreover, most of the current detection methods only detect the height of the mold liquid level and its flow rate, and use Eddy current to detect the electromagnetic induction of the metal. However, such detection can only detect speed (because electromagnetic waves are related to the speed of current flow), and only have the ability to detect a single point. Because the detection position of the above detection method is fixed, but the location where the explosion occurs will be everywhere in the mold liquid surface, it is impossible to accurately determine the location where the explosion will occur on the mold liquid surface, and thus the sudden explosion boil cannot be detected. The problem of slag entrapment caused by etc. cannot solve the problem of steel defects.

Furthermore, the existing use of thermal images to observe surface changes is done by human eyes, so it is impossible to quantify the data and provide automated monitoring functions.

Therefore, it is necessary to provide a method for detecting liquid level disturbance in surface image to solve the problems of conventional technology.

The main purpose of the present invention is to provide a surface image liquid level disturbance detection method, which can accurately determine the position where a sudden explosion occurs on the mold liquid surface.

The secondary purpose of the present invention is to provide a surface image liquid level disturbance detection method, which uses the image value statistical calculation method of the thermal image to achieve a large range of boiling detection in a surface manner, and uses the same area of the image The difference between the front and rear frames is used to detect the area and time of rolling boiling by the degree of value change of the difference value. The calculation of the difference value is proportional to the degree of change of the mold liquid level, which can replace the human eye to observe the size of the mold liquid level change. Detection.

The secondary objective of the present invention is to provide a surface image liquid level disturbance detection method, which uses thermal image as a sensing method, instead of the existing single-point detection method of eddy current, since thermal image can be used for remote sensing, installation It is more convenient than other methods, and the image sensing itself is a two-dimensional planar measurement, which can measure planar information more than other sensing methods. Furthermore, the real-time streaming of thermal images and the calculation of the regional changes of the continuous frame of thermal images can quantify the changes in thermal images to achieve the numerical goal. Using numerical analysis again, it is possible to find out the time and area of the image where the value of the value change is too high, and summarize the state of the value change of boiling, to achieve the effect of automatic detection and warning.

To achieve the above objective, an embodiment of the present invention provides a surface image liquid level disturbance detection method, which includes the following steps: (a) providing a mold liquid level during continuous heating; (b) providing a continuous mold liquid level A first frame, a second frame, and a third frame in several time intervals before and after; (c) defining a first frame around the first area of each of the first frame, the second frame, and the third frame Two areas, where the second area is larger than the first area; (d) in the second area between the first frame, search for an area similar to the image in the first area of the second frame to calculate a A first degree of similarity; and (e) after replacing the first degree of similarity with a degree of difference, a roll boiling probability is obtained to determine a roll boiling degree in the second area of the first frame.

In an embodiment of the present invention, the mold liquid level is a mold liquid level continuously cast in a steelmaking plant.

In an embodiment of the present invention, the surface image liquid level disturbance detection method further includes: continuously monitoring the mold liquid level with a thermal imager.

, In an embodiment of the present invention, the rolling boiling probability is expressed by the following formula:

,

表示該第一畫幅在該第二畫幅中的該第二區域的一相似度；

表示該第一畫幅在該第二畫幅中的該第二區域的一差異度；及ℝ表示該第二區域內的範圍。 among them

Represents a similarity of the first frame in the second area in the second frame;

Represents a degree of difference between the first frame and the second area in the second frame; and ℝ represents the range within the second area.

In an embodiment of the present invention, step (d) further includes: in the second area of the third frame, searching for an area similar to the image in the first area of the second frame to calculate a A second degree of similarity; and after converting the second degree of similarity into a degree of difference, a rolling probability is obtained to determine the degree of rolling in the second region of the third frame.

In an embodiment of the present invention, the surface image liquid level disturbance detection method further includes: (f) defining a first frame around a third area of each of the first frame, the second frame, and the third frame Four areas, where the fourth area is larger than the third area; (g) in the fourth area between the first frame, search for an area similar to the image in the third area of the second frame to calculate a A third degree of similarity; (h) in the fourth area between the third frame, search for an area similar to the image in the third area of the second frame to calculate a fourth degree of similarity; and (i) After converting the third degree of similarity and the fourth degree of similarity to another degree of difference, another probability of rolling boiling is obtained to determine another degree of rolling boiling in the fourth region, and then the entire liquid surface of the mold The surface is monitored for rolling boiling.

In an embodiment of the present invention, the third area does not overlap with the first area; and the fourth area does not overlap with the second area.

In an embodiment of the present invention, the surface image liquid level disturbance detection method further includes: marking the area with a high degree of boiling.

In order to make the above and other objectives, features, and advantages of the present invention more obvious and understandable, the following will specifically cite the preferred embodiments of the present invention, together with the accompanying drawings, and describe in detail as follows. Furthermore, the directional terms mentioned in the present invention, such as up, down, top, bottom, front, back, left, right, inside, outside, side, surrounding, center, horizontal, horizontal, vertical, vertical, axial, The radial direction, the uppermost layer or the lowermost layer, etc., are only the direction of reference to the attached drawings. Therefore, the directional terms used are used to describe and understand the present invention, rather than to limit the present invention.

Please refer to Figure 1, which shows a surface image liquid level disturbance detection method according to an embodiment of the present invention, including the following steps: (S11) providing a mold liquid level in continuous heating; (S12) providing the A first frame, a second frame, and a third frame in several time intervals before and after the mold liquid level is continuous; (S13) in each of the first frame, the second frame, and the third frame A second area is defined around the area, wherein the second area is larger than the first area; (S14) in the second area between the first frames, search for images similar to the images in the first area of the second frame Area to calculate a first degree of similarity; and (S15) after replacing the first degree of similarity with a degree of difference, a rolling probability is obtained to determine the degree of rolling in the second region of the first frame .

First, please refer to step S11 to provide a mold liquid level during continuous heating. In this embodiment, the mold liquid level is a mold liquid level continuously cast in a steelmaking plant. A thermal imager is arranged on an upper surface of the mold liquid level to perform a thermal image monitoring on the mold liquid level.

Next, referring to step S12, provide a first frame, a second frame, and a third frame in several time intervals before and after the mold liquid level is continuous. Optionally, the method for detecting surface image liquid level disturbance further includes: continuously monitoring the mold liquid level with a thermal imager. The thermal imager continuously monitors a thermal image data of the mold liquid level. Optionally, the thermal imager is connected to a storage device to store the thermal image data in the storage device. The thermal image data is a series of continuous time images, the color depth of which corresponds to different temperature values. In this embodiment, the thermal image stream includes the first frame, the second frame, and the third frame in several time intervals before and after the mold liquid level is continuous.

…式1 The rolling boiling probability of the mold liquid surface can be expressed by the following formula 1:

…Formula 1

代表在某時刻

在某時刻x之滾沸機率，其正比於溫度T的變化。當模液面熱影像發生的變動越大，也代表其滾沸的機率越高。 among them

Represents at a certain moment

The rolling boiling probability of x at a certain time is proportional to the change of temperature T. The greater the change in the thermal image of the mold liquid level, the higher the probability of boiling.

…式2 To measure the rolling boiling probability in a region, you can integrate the above formula for a region, as shown in the following formula 2:

…Form 2

表示一特定的區域。 among them,

Represents a specific area.

In order to compare the difference value of a specific area, time can be differentiated. The above-mentioned differentiation can use high-pass filtering to achieve the effect of differentiation. Compare the first frame, the second frame, and the third frame in the first and second time intervals in the image data with respect to the image in the first area and the before and after frames to obtain the image data in the first area The difference value.

或

Please refer to FIG. 2 to perform difference numerical analysis on the images in the first area of the first frame, the second frame, and the third frame. The above-mentioned difference numerical analysis can be expressed by the following formula:

or

…式3 According to the numerical analysis of the difference in the above formula, the above formula 2 can be rewritten as the following formula 3:

…Formula 3

The degree of rolling boiling in the first zone within a certain temperature range can be proportional to the following formula:

表示在時間t時在x位置的溫度變化。而

可以解讀為連續的數個時間區間內的該第一畫幅、該第二畫幅及該第三畫幅中的

區域的相似程度。影像相似程度越高（差異性越小）代表其滾沸機率越低。 among them

Represents the temperature change at position x at time t. and

Can be interpreted as the first frame, the second frame, and the third frame in several consecutive time intervals

The similarity of the regions. The higher the degree of image similarity (the smaller the difference), the lower the probability of boiling.

。 Next, referring to step S13, a second area is defined around the first area of each of the first frame, the second frame, and the third frame, wherein the second area is larger than the first area. In other words, define the second area as

.

區域內搜索與該第二畫幅的該第一區域中影像相似的區域，並判斷其相似程度。 Referring to step S14 and shown in Figure 3, in the second area between the first frame, search for an area similar to the image in the first area of the second frame to calculate a first similarity. In other words, based on the image in the first area of the second frame, the

Search for an area similar to the image in the first area of the second frame in the area, and determine the degree of similarity.

Next, referring to step S15 and shown in FIG. 3, after the first similarity is replaced by a difference degree, a rolling boiling probability is further obtained, so as to determine a rolling boiling degree in the second region of the first frame.

Optionally, the method for detecting liquid level disturbance of the surface image, wherein the step (S14) further includes: searching for an image in the first area of the second frame in the second area of the third frame Similar regions to calculate a second degree of similarity; and after converting the second degree of similarity to a degree of difference, a rolling probability is obtained to determine the degree of rolling in the second region of the third frame. In this way, the n frames are compared with the similarity of the second region in the manner described above.

…式4 Searching for an area similar to the image in the first area of the first frame in the second area of the front and rear frames, the above formula 3 can be rewritten as the following formula 4:

…Formula 4

表示該第一畫幅在該第二畫幅中的該第二區域的一相似度；

表示該第一畫幅在該第二畫幅中的該第二區域的一差異度；及ℝ表示該第二區域內的範圍。 among them

Represents a similarity of the first frame in the second area in the second frame;

Represents a degree of difference between the first frame and the second area in the second frame; and ℝ represents the range within the second area.

The above calculation method can be as shown in Figure 4, where the P interval (for example, the first area) of one frame (for example, the second frame) is in the R interval of another frame (for example, the first frame) (For example, the second area), search for the area with the greatest similarity. The above calculation method can be used to determine images in multiple consecutive time segments. Optionally, the surface image liquid level disturbance detection method further includes: (S16) defining a fourth area around a third area of each of the first frame, the second frame, and the third frame, wherein the The fourth area is larger than the third area; (S17) in the fourth area between the first frames, search for an area similar to the image in the third area of the second frame to calculate a third similarity; (S18) in the fourth area between the third frame, search for an area similar to the image in the third area of the second frame to calculate a fourth similarity; and (S19) the third similarity After the degree of similarity and the fourth degree of similarity are converted into another degree of difference, another rolling boiling probability is obtained to determine another rolling boiling degree in the fourth region, and then rolling boiling monitoring is performed on the entire liquid surface of the mold liquid surface.

Preferably, the third area does not overlap with the first area; and the fourth area does not overlap with the second area.

Finally, referring to step S16, after changing the first degree of similarity to a degree of difference, a rolling boiling probability is further obtained to determine a rolling boiling degree in the second region.

Optionally, the surface image liquid level disturbance detection method further includes: marking the area with a high degree of boiling.

Please refer to Figure 5, after continuous comparison and continuous calculation processing, continuous time rolling data can be obtained.

After numerical analysis and processing of the continuous-time rolling boiling data obtained in Figure 5, the rolling boiling probability can be converted to as shown in Figure 6, where the left side (such as the first frame) represents a lower rolling boiling probability, and the right (such as the third frame) Represents a higher probability of rolling boiling. In other words, the second area of the third frame has a tendency to increase the probability of boiling.

Since rolling boiling is a continuous phenomenon, and the disturbance of the liquid level during rolling boiling is mostly maintained for a period of time, the embodiments of the present invention can use the numerical value and the high frequency of occurrence to warn whether there is rolling boiling and supply the rear process system For correspondence.

Since this calculation method involves the plane calculation of the area, when the detection interval is too large, the influence of the area change is too small, making the numerical difference not enough to highlight the boiling state; on the contrary, when the detection interval is too small, small changes will cause great The value changes, making the system overly sensitive. So the size of the comparison interval will affect the sensitivity. Therefore, the embodiments of the present invention can adjust the size of the detection interval according to requirements. When a large-scale operation detection is required, it is necessary to cover the operation area with the same size operation block for independent operation to obtain the rolling value of each local area. Avoid using too large arithmetic area to perform calculations and reduce the sensitivity to roll boiling changes.

The present invention uses continuous time thermal images to detect boiling in a specific area. The calculation of the degree of rolling is obtained by using the continuous time difference of the images of a specific area. The calculation of the difference can be obtained by the degree of difference between the image values of the two regions. The embodiment of the present invention can increase the resistance to the disturbance of the liquid level, thereby increasing the search area, thereby obtaining the largest similar area in the area as the calculated value of the rolling boiling probability, and reducing the misjudgment caused by the disturbance. The embodiment of the present invention can cover the surface of the mold liquid surface with multiple calculation areas, and perform the boiling detection of the entire liquid surface. Since each operation area is an independent operation area, it is suitable for parallel operation.

The above-mentioned embodiments only illustrate the principles and effects of the present invention, and do not limit the present invention. Therefore, those who are accustomed to this technology can modify and change the above-mentioned embodiments without departing from the spirit of the present invention. The scope of rights of the present invention should be listed in the scope of patent application described later.

S11～S15: steps

[Figure 1]: A schematic flow diagram of a method for detecting liquid level disturbance in a surface image according to an embodiment of the present invention; [Figure 2]: A schematic diagram of a method for detecting liquid level disturbances in a surface image according to an embodiment of the present invention for numerical analysis of differences in images of the first region; [Figure 3]: A schematic diagram of the surface image liquid level disturbance detection method of an embodiment of the present invention searching for an area similar to an image in the first area of the first frame in a second area of a different frame; [Figure 4]: A schematic diagram of the calculation method of the surface image liquid level disturbance detection method according to an embodiment of the present invention; [Figure 5]: A schematic diagram of continuous time rolling data obtained after continuous comparison and continuous calculation processing of the surface image liquid level disturbance detection method of an embodiment of the present invention [Figure 6]: A schematic diagram of the surface image liquid level disturbance detection method of an embodiment of the present invention converted into a rolling boiling degree after numerical analysis processing.

S11~S15: steps

Claims (8)

A method for detecting liquid level disturbance in a surface image includes the following steps: (a) Provide a mold liquid level during continuous heating; (b) Provide a first frame, a second frame, and a third frame in several time intervals before and after the mold liquid level is continuous; (c) defining a second area around the first area of each of the first frame, the second frame, and the third frame, wherein the second area is larger than the first area; (d) in the second area between the first frame, search for an area similar to the image in the first area of the second frame to calculate a first similarity; and (e) After changing the first degree of similarity to a degree of difference, a rolling boiling probability is obtained to determine the rolling boiling degree in the second area of the first frame. According to the method for detecting surface image liquid level disturbance described in item 1 of the scope of patent application, the mold liquid level is a mold liquid level of continuous casting in a steelmaking plant. The surface image liquid level disturbance detection method as described in claim 1, wherein the surface image liquid level disturbance detection method further includes: continuously monitoring the mold liquid level with a thermal imager. According to the method for detecting liquid level disturbance in surface image as described in item 1 of the patent application, the rolling boiling probability is expressed by the following formula:

, among them

Represents a similarity of the first frame in the second area in the second frame;

Represents a degree of difference between the first frame and the second area in the second frame; and ℝ represents the range within the second area. The surface image liquid level disturbance detection method described in claim 1, wherein step (d) further comprises: searching for the second area of the third frame and the second area of the second frame The image similar area in the first area is used to calculate a second degree of similarity; and after the second degree of similarity is converted into a degree of difference, a rolling probability is obtained to determine the second area in the third frame A degree of boiling. The surface image liquid level disturbance detection method as described in item 1 of the scope of patent application, wherein the surface image liquid level disturbance detection method further includes: (f) defining a fourth area around a third area of each of the first frame, the second frame and the third frame, wherein the fourth area is larger than the third area; (g) In the fourth area between the first frame, search for an area similar to the image in the third area of the second frame to calculate a third degree of similarity; (h) In the fourth area between the third frame, search for an area similar to the image in the third area of the second frame to calculate a fourth similarity; and (i) After converting the third degree of similarity and the fourth degree of similarity to another degree of difference, another probability of rolling boiling is obtained, so as to determine another rolling boiling degree in the fourth region, and then the mold liquid level The entire liquid level is monitored for rolling boiling. According to the method for detecting liquid level disturbance in a surface image according to claim 7, wherein the third area does not overlap with the first area; and the fourth area does not overlap with the second area. The surface image liquid level disturbance detection method as described in item 1 of the scope of patent application, wherein the surface image liquid level disturbance detection method further comprises: marking the area with a high degree of boiling.

Images