TWI700567B - Image analysis apparatus for industrial plants and industrial plant monitoring control system - Google Patents

Abstract

An image analysis apparatus for industrial plants and an industrial plant monitoring control system have the following configurations. A moving image data collecting unit (21) collects moving image data obtained by shooting an equipment constituting an industrial plant and materials to be processed by the equipment in real time. An image processing treatment unit (22) extracts an image from the moving image data for each constant period, and binarizes the image by converting specified color into first color while converting color other than the specified color into second color. An image quantifying unit (23) quantifies the binarized image on the basis of the number of pixels converted into the first color. A numerical value output unit (24) outputs the quantified data.

Description

Image analysis device for industrial factory and industrial factory monitoring and control system

The present invention relates to an image analysis device for an industrial factory and an industrial factory monitoring and control system.

Known to have industrial plants (steel plants, power plants, petroleum plants, chemical plants, etc.) that produce materials and resources required for industrial activities. The factory monitoring and control system of an industrial factory is equipped with input and output devices (I/O) connected to a large number of field devices (including actuators and sensors) that constitute the factory through a control network, and can control a large number of field devices. A configuration in which programming logic controllers (hereinafter referred to as PLCs) are connected to each other.

In addition, the process data as the input and output signals of the PLC and I/O devices are collected in a data collection device with data collection and data playback functions. The data collection device displays the collected process data and is used by the operator to grasp the status of the industrial plant.

As for an example of a sensor that outputs process data, Patent Document 1 has disclosed an HMD (Hot Metal Detector) installed in a steel factory. The HMD is installed at a predetermined height position directly above the roller table downstream of the roll, and the detection direction is set in a direction substantially perpendicular to the roll and facing the horizontal direction. HMD is a laser sensor with a narrow field of view. If thermal mass is detected in the field of view, it outputs an ON signal, and if no thermal mass is detected, it outputs an OFF signal.

[Prior Technical Literature] [Patent Literature]

Patent Document 1: Japanese Patent Application Publication No. 2016-87652

Patent Document 2: International Publication No. 2014/002176

As described in Patent Document 1, the upward warpage of the thermal mass (rolled material) can be detected by disposing a sensor. However, the sensor is expensive and its installation location is also limited. In addition, the sensor only outputs an ON signal or an OFF signal, and cannot grasp the state of the rolled material (shape, position, etc.) in detail.

In order to solve the above-mentioned shortcomings, the inventor of the present case has conducted in-depth research and developed the dynamic image data of the subject taken with a general camera to be used as a sensor. By using a camera, the advantages of cost reduction, wider shooting range compared with the above-mentioned sensor, higher freedom of installation location, and more information output than the sensor can be obtained.

On the other hand, when confirming the state of the subject photographed by the camera, the visual observation of the dynamic image data is only a visual judgment, and the accuracy of the confirmation is low. Therefore, processing to improve the accuracy of confirmation is desired.

The present invention was made in order to solve the above problems, and its purpose is to provide an image analysis device for industrial factories, which uses dynamic image data to quantitatively analyze the state of the subject in real time, and can support operators who confirm the operating conditions of industrial factories .

In addition, another object of the present invention is to provide an industrial factory monitoring and control system that uses data quantified by the industrial factory's image analysis device to achieve appropriate control according to the state of the subject.

In order to achieve the above-mentioned object, the image analysis device for industrial factories of this embodiment is configured as follows. The image analysis device for industrial plants is equipped with a dynamic image data collection unit, an image processing unit, an image quantization unit, and a numerical output unit.

The dynamic image data collection department collects real-time dynamic image data obtained by photographing the machinery that constitutes an industrial factory and the materials processed in the machinery. Real-time does not necessarily refer to the moment of shooting. It includes data collection with delays due to communication paths and data processing, or data collection at a certain interval.

The image processing unit extracts an image from the aforementioned dynamic image data every certain period, converts the specified color into a first color, and converts colors other than the specified color into a second color, and binarizes the image. The certain period is preferably as short as possible, and more preferably instant. Portrait refers to a frame of portrait extracted from dynamic portrait data. The specified color is not only a single color, but also includes the specified color range.

The image quantization unit quantifies the binarized image based on the number of pixels converted into the first color. In one aspect, the image quantization unit calculates the proportion of the first color in each band-shaped region formed by dividing the binarized image parallel to the conveying direction of the material. In another aspect, the image quantization unit calculates the proportion of the first color in each grid-like region that divides the binarized image into a grid.

The numerical output unit outputs the aforementioned quantized data.

To achieve the above-mentioned object, the industrial plant monitoring and control system of this embodiment is configured as follows. The industrial factory monitoring and control system is equipped with the aforementioned industrial factory image analysis device and a programmable logic controller that controls the aforementioned equipment.

The aforementioned machine system includes a pair of rolls for rolling a rolled material as the aforementioned material. The image processing unit binarizes the image of the rolling material viewed from the width direction.

In one aspect, the image quantization unit and the programmable logic controller are configured as follows.

The image quantization unit makes the binarized image parallel to the conveying direction of the rolling material, and divides it into at least the first area and the second area adjacent to the first area, and calculates the amount of the first color in each area proportion.

The programmable logic controller outputs a warning when the difference between the increase in the proportion of the first color in the second area and the decrease in the proportion of the first color in the first area is greater than the threshold At least one of a signal and a control signal. The control signal changes the rotation speed of the pair of rolls to suppress the upward warping of the rolled material. For example, the control signal is a signal for stopping the aforementioned pair of rolls, or a signal for making the rotation speed of the lower roll slower than the rotation speed of the upper roll.

In another aspect, the image quantization unit and the programmable logic controller are configured as follows.

The image quantization unit divides the binarized image into at least a grid-like first area, a second area adjacent to the first area laterally, a third area adjacent to the upper side of the first area, and a second area adjacent to the second area. The fourth area above the area and adjacent to the third area laterally, and the proportion of the aforementioned first color in each area is calculated.

In the programmable logic controller, the proportion of the aforementioned first color increases in the order of the first area and the second area, and the proportion of the aforementioned first color in the fourth area is based on the aforementioned first color in the third area. When the proportion of the color decreases and increases, at least one of a warning signal and a control signal is output. The control signal changes the rotation speed of the pair of rollers to suppress the upward warpage of the rolled material. For example, the control signal is a signal to stop the aforementioned pair of rolls or a signal to make the rotation speed of the lower roll slower than the rotation speed of the upper roll.

According to the image analysis device for industrial factories of the present invention, dynamic image data can be used to quantitatively analyze the state of the subject in real time, and can support operators who confirm the operating conditions of industrial factories. Furthermore, according to the industrial factory monitoring and control system of the present invention, appropriate control can be achieved by using data quantified by the image analysis device for the industrial factory.

1‧‧‧Image analysis device

2‧‧‧The first surveillance camera

3‧‧‧Second surveillance camera

4‧‧‧Image data converter

5‧‧‧Signal line for portrait

6‧‧‧Network for dynamic portrait

7‧‧‧Data collection device

8‧‧‧Human Machine Interface

9‧‧‧Programmable Logic Controller

10‧‧‧Control network

21‧‧‧Dynamic Portrait Data Collection Department

22‧‧‧Image Processing Department

23‧‧‧Image Quantization Department

24‧‧‧Numerical output

31‧‧‧Left Window

32‧‧‧Central Window

33‧‧‧Right window

41‧‧‧Color filter

42‧‧‧Binaryization Department

43‧‧‧Image Processing Department

50‧‧‧Strip quantization processing section

51‧‧‧Band A

52‧‧‧Band B

53‧‧‧belt area C

54‧‧‧Banded area D

60‧‧‧Grid-shaped quantization processing unit

61‧‧‧Grid-shaped area A

62‧‧‧Grid-shaped area B

63‧‧‧Grid-like area C

64‧‧‧Grid-like area D

65‧‧‧Grid-like area E

66‧‧‧Grid-like area F

67‧‧‧Grid-shaped area G

68‧‧‧Grid-like area H

69‧‧‧Grid-like area I

70‧‧‧Grid-like area J

71‧‧‧Grid-like area K

72‧‧‧Lattice area L

111‧‧‧CPU

112‧‧‧Memory

113‧‧‧Storage

113a‧‧‧Program Memory

113b‧‧‧Data Memory Department

114‧‧‧External machine interface

115a‧‧‧Network interface for control

115b‧‧‧Network interface for dynamic image

116‧‧‧Internal bus

117‧‧‧Display

118‧‧‧Keyboard

119‧‧‧Mouse

311‧‧‧Rolled products

312‧‧‧roll

313‧‧‧Transport roller

321‧‧‧ indicates the part of rolled material

322‧‧‧Other parts

Fig. 1 is a schematic diagram showing the configuration of the factory monitoring and control system of the industrial factory in the first embodiment.

Fig. 2 is a block diagram of the image analysis device in the first embodiment.

Fig. 3 is an example of screen display of the image analysis device of the first embodiment.

Figure 4 is a block diagram showing an example of the hardware configuration of the processing circuit included in the image analysis device.

Fig. 5 is a block diagram of the image analysis device of the second embodiment.

Fig. 6 is an analysis example of the image analysis device of the second embodiment.

Fig. 7 is a block diagram of the image analysis device of the third embodiment.

Fig. 8 is an analysis example of the image analysis device in the third embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, the same reference numerals are given to common elements in the respective drawings, and repeated descriptions are omitted.

Embodiment 1

(System Components)

Fig. 1 is a schematic diagram showing the configuration of the factory monitoring and control system of the industrial factory in the first embodiment.

A steel factory is an industrial factory that produces materials and resources required for industrial activities. In hot rolling, the high-temperature slab (rolled material) heated by the heating furnace is transported by conveying rollers, stretched thinly to the required thickness by the roughing mill and finishing mill, and processed to the required width at the same time, and finally used Coiler winding.

The factory monitoring and control system is connected to the image analysis device 1, the first monitoring camera 2, the second monitoring camera 3, the data collection device 7, the HMI (HUMAN MACHINE INTERFACE, human-machine interface) 8, the PLC 9 and the machines that constitute the industrial factory (omitted Illustrated) and so on.

The first monitoring camera 2 and the second monitoring camera 3 are arranged so that the equipment constituting the industrial factory and the materials processed in the machine are included in the shooting range. The first monitoring camera 2 is connected to the image document converter 4 via the image signal line 5. The image output of the first monitoring camera 2 is converted by the image data converter 4 into a signal capable of network communication, and is transmitted to the image analysis device 1 via the dynamic image network 6. The second monitoring camera 3 is a network camera, and the captured dynamic image data is transmitted to the image analysis device 1 via the dynamic image network 6.

The image analysis device 1 is connected to the programmable logic controller (PLC) 9 that controls the machines (including actuators and sensors) constituting the industrial factory via the control network 10, and is used as an operator for general operations in the industrial factory or The human-machine interface (HMI) 8 and the data collection device 7 of the monitoring control device used in the monitoring. In addition, there are cases where the image analysis device 1 is incorporated into the data collection device 7.

The control network 10 is provided with a plurality of nodes with common memory, and the data on the common memory is synchronized by periodic broadcast transmission between the plurality of nodes. In this way, the same virtual memory space is shared between the image analysis device 1, the PLC 9, and the HMI 8. The memory area (address) of each data is allocated in the common memory. The devices connected to each node can send and receive data by writing to and reading from the public memory.

(Image analysis device)

The image analysis device 1 includes a dynamic image data collection unit 21, an image processing unit 22, an image quantization unit 23, and a numerical output unit 24.

The dynamic image data collection unit 21 collects and photographs the dynamic image data of the machines constituting the industrial factory and the materials processed in the machines in real time. Real-time does not necessarily refer to the moment of shooting, but also includes data collection accompanied by delays due to communication paths and data processing, or data collection at regular intervals. Specifically, the dynamic image data collection unit 21 collects dynamic image data from the first surveillance camera 2 and the second surveillance camera 3 via the dynamic image network 6 at regular intervals. The dynamic image data collection unit 21 outputs the dynamic image data to the image processing unit 22. The dynamic image data collection unit 21 stores the collected dynamic image in the data storage unit 113b of the memory 113.

The image processing unit 22 extracts an image from the dynamic image data every certain period, converts a specified color into a first color, and converts colors other than the specified color into a second color, and binarizes the image. The certain period is preferably as short as possible, and more preferably instant. Portrait refers to a frame of portrait extracted from dynamic portrait data. The specified color is not only a single color, but also includes the specified color range.

Specifically, the image processing unit 22 processes the image using the color filter unit 41, the binarization unit 42, and the image processing unit 43 shown in FIG. 2, and then outputs the processed image to the image quantization unit 23.

The color filter unit 41 extracts the designated colors related to the equipment and materials in the image. As for the designated color, for example, the color of hot material, the color of the abnormal temperature of the display machine or material, etc. are set in advance.

The binarization unit 42 converts the designated color into a first color (for example, white), and converts colors other than the designated color into a second color (for example, black), and binarizes the image. This can reduce the amount of data, reduce the computational load of image processing and quantification processing, and ensure real-time.

When the first monitoring camera 2 or the second monitoring camera 3 captures the subject diagonally, the image processing unit 43 performs image conversion to convert it into an image viewed from the side or from above. In this way, the flexibility of camera installation is high, not only in situations where the camera cannot be installed above or beside the subject, but also in environments where it is difficult to install a sensor near the subject. In addition. When a camera can be installed above or beside the subject, it is not necessary to execute the processing of the image processing unit 43.

The image quantization unit 23 quantifies the aforementioned binary image based on the number of pixels converted into the first color. For example, the ratio of the first color in the image is calculated based on the number of pixels converted into the first color (for example, white) by the image processing unit 22. The image quantization unit 23 outputs the quantized data to the numerical output unit 24.

The numerical output unit 24 outputs the quantified data to the data collection device 7, the HMI 8, and the PLC 9 via the control network 10.

FIG. 3 is a diagram showing a display example displayed by the image analysis device 1 on the display 117 (refer to FIG. 4). The image analysis device 1 displays the collected dynamic image data in the left window 31 in real time. In the example in Fig. 3, it is shown that the shooting range includes a pair of rolls 312 (roughing mill or finishing mill) and conveying rollers 313 of the rolled material 311 that is rolled as the material. The rolled material 311 is conveyed from left to right by a pair The rolling state of the roll 312. In addition, the image analysis device 1 displays an image in which the image displayed in the left window 31 is binarized in the center window 32. The part 321 representing the rolled material is displayed in white, and the other part 322 is displayed in black. In addition, the image analysis device 1 displays a graph showing the temporal change of the proportion of white in the binarized image in the right window 33. In the example of Fig. 6, a graph showing the time change of the proportion of white in each area divided in the horizontal direction is displayed.

The data collection device 7 collects process data from the PLC 9 and HMI 8. Process data includes various data related to the machines that constitute the industrial plant and the materials processed by the machines. For example, it includes the control value of the actuator, the detection value of the sensor, and the material specification. In large-scale factories such as steel plants, there are thousands or tens of thousands of input and output points, and there are various process data. The process data is collected in the data collection device 7 with data collection function and data playback function, and used for data analysis during testing, adjustment, operation, and failure.

In addition, the data collection device 7 receives quantified data from the image analysis device 1 and synchronizes with the process data and collects and stores them simultaneously. In addition, the data collection device 7 receives necessary data from the image analysis device 1, and can realize the same display as in FIG. 3.

HMI 8 displays the process data received from PLC 9 in numerical value, text and light. In addition, HMI 8 is equipped with operation buttons for sending button input signals and numerical input signals, and outputs signals for controlling PLC 9 according to pressing the operation buttons. In addition, the HMI 8 receives quantified data from the image analysis device 1, and can display its value, or can issue an alarm when its value exceeds a predetermined threshold.

The PLC 9 performs calculations based on the input of sensors from industrial factories and outputs signals to actuators such as valves and motors to control industrial factories. In addition, the data output from the image analysis device 1 is received in the same manner as the input from the sensor to perform calculations, and signals are output to actuators such as valves and motors to control industrial plants. In addition, an alarm is issued when the calculation result exceeds the critical value.

As described above, according to the image analysis device 1 of the first embodiment, by quantifying the dynamic image collected from the surveillance camera in real time, it is possible to grasp the extent to which the subject has entered the shooting range. Therefore, the image analysis device 1 can quantitatively analyze the position and shape of the machine and material in real time, and can support the operator to confirm the operation status. In addition, since no special sensor is used, the cost is lower. For example, it is possible to reduce the installation cost of sensors for detecting warpage of hot-rolled steel strips and thick plates. In addition, compared with the sensor, the shooting range is wide and the degree of freedom of the installation location is high, so the state of the subject can be grasped in various environments.

In the first embodiment described above, the number of pixels converted into the first color is calculated for the entire image, but it is not limited to this. You can also specify a partial area of the portrait, and calculate the number of pixels in the area converted to the first color. By specifying the area that should be paid attention to in the image, noise can be reduced and detection accuracy can be improved.

In addition, by using the image analysis device 1 and the data collection device 7 of the first embodiment, the dynamic image data can be displayed on the data collection device 7 in real time, and the data stored in the data collection device 7 can also be played back from any specified time. Collect the dynamic image data of the device 7. It is easy to grasp the position and size of the past phenomenon, and the confirmation accuracy of the position and size can be improved. In addition, you can also stop, fast forward, and rewind the dynamic image. In addition, the above content is the same in the following embodiments.

In addition, although the image analysis device 1 of the first embodiment described above does not include the display 117, the keyboard 118, and the mouse 119 shown in FIG. 4 described later, the image analysis device 1 may include these equipment. In addition, the above content is the same in the following embodiments.

In addition, although the industrial system was explained by taking a steel factory as an example in the above-mentioned Embodiment 1, it is not limited to this. The industrial system also includes power plants, petroleum plants, chemical plants, etc.

(Example of hardware configuration)

The hardware configuration of the image analysis device 1 will be described with reference to Fig. 4. FIG. 4 is a block diagram showing an example of the hardware configuration of the processing circuit included in the image analysis device 1. Each part of the image analysis device 1 shown in FIG. 2 represents a part of the functions of the image analysis device 1, and each function is realized by a processing circuit. For example, the processing circuit is connected to the CPU 111, the memory 112, the storage 113 such as HDD or large-capacity memory, the external machine interface (I/F) section 114, the control network interface section 115a, and the dynamic image via the internal bus 116 It is constituted by the network interface part 115b. In addition, the data collection device 7 has the same configuration.

The CPU 111 implements the functions of each part of the image analysis device 1 by executing various programs stored in the program memory 113a of the memory 113.

The memory 112 is used as a calculation area for temporarily storing or expanding data when the CPU 111 executes various programs.

The storage 113 has a program storage unit 113a and a data storage unit 113b. The program memory 113a stores OS (OPERATING SYSTEM) and various programs. In addition, the data storage unit 113b memorizes the collected process data and dynamic image data at each time.

Furthermore, in the example shown in FIG. 4, although the program storage unit 113a and the data storage unit 113b are provided in one storage 113, the program storage unit 113a and the data storage unit 113b may be separately provided in a plurality of storages.

The external device interface unit 114 is an interface for connecting external devices such as the display 117, the keyboard 118, and the mouse 119 to the image analysis device 1.

The control network interface unit 115a is an interface for connecting the control network 10 and the image analysis device 1. In addition, the dynamic image network interface unit 115b is an interface for connecting the dynamic image network 6 and the image analysis device 1.

Embodiment 2

Next, the second embodiment will be described with reference to Figs. 5 and 6. In the first embodiment described above, the ratio of the first color to the overall image is calculated. On the other hand, in the second embodiment, the proportion of the first color in each band-shaped region that divides the image into bands is calculated, and the change in the proportion of the first color in each band-shaped region is compared in more detail. Analyze the state of the subject.

Fig. 5 is a block diagram of the image analysis apparatus 1 of the second embodiment. The configuration shown in FIG. 5 is the same as that shown in FIG. 2 except that the image quantization unit 23 includes a band-shaped quantization processing unit 50, so the description of the common configuration is omitted or simplified.

The image processing unit 22 binarizes the image of the material viewed from the width direction.

The band-shaped quantization processing unit 50 divides the image binarized by the image processing unit 22 into each band-shaped area parallel to the conveying direction of the material, and calculates the first color (for example: white) occupied proportion. In the example in Figure 5, the binarized image system is divided into 4 regions in the horizontal direction. The four divided regions are referred to from bottom to top as a band-shaped region A51, a band-shaped region B52, a band-shaped region C53, and a band-shaped region D54. The band quantization processing unit 50 calculates the ratio of the number of white pixels in each band region.

Calculate the ratio of the first color in each band-shaped area (hereinafter referred to as the first color ratio) by dividing the image into layers in real time, and by confirming the change in the first color ratio, you can quantitatively grasp the position of the subject , Changes in shape.

For example, the rolled material 311 (refer to FIG. 3) used as a material may jump on the conveying roller 313 (refer to FIG. 3) of a steel factory. At this time, the photographic subject (rolled material) moves in parallel from the band-shaped area A51 to the band-shaped area B52. Therefore, by confirming that the increase in the first color ratio of the belt-shaped area B52 is equal to the decrease of the first color ratio in the belt-shaped area A51, it can be confirmed that the photographic subject has moved in parallel.

On the other hand, the shape of the subject may be deformed. For example, in a hot rolling line and a thick plate line of a steel plant, the rolled material 311 (refer to Fig. 3) may be warped upward due to rolling. Refer to Figure 6 to explain an example of image analysis. t1 is the shape of the rolled material in the previous cycle, and t2 is the shape of the rolled material in the current cycle. The band-shaped quantization processing unit 50 divides the binarized image into at least a band-shaped area A51 and a band-shaped area B52 adjacent to the upper side of the band-shaped area A51 parallel to the conveying direction of the rolled material, and calculates the first A color ratio. The boundary between the band-shaped area A51 and the band-shaped area B52 is set between the upper surface and the lower surface of the rolling material.

The operator confirms the change in the first color ratio in each band-shaped area accompanying the change from t1 to t2. By confirming the magnitude of the difference between the increase in the first color ratio in the band-shaped area B52 and the decrease in the first color ratio in the band-shaped area A51, the upward warpage of the rolled material can be quantitatively grasped.

The image analysis device 1 can support the operator's confirmation operation by displaying a graph showing the time change of the first color ratio in each band-shaped area as illustrated in the right window 33 in FIG. 3.

The PLC9 receives the data (the first color ratio in each band-shaped area) output by the numerical output unit 24. The PLC9 can perform control based on the position of the subject, control based on the amount of movement of the subject, and control based on the shape of the subject from the first color ratio in each band-shaped area (for example, control based on the amount of upward warpage) .

The control based on the amount of upward warpage will be described with reference to the above-mentioned Fig. 6. The PLC9 outputs at least one of a warning signal and a control signal when the difference between the increase in the first color ratio in the belt-shaped area B52 and the decrease in the first color ratio in the belt-shaped area A51 is greater than the critical value, and the control The signal is to change the rotation speed of the pair of rollers 312 to suppress the upward warpage of the rolled material. For example, the control signal is a signal to stop the pair of rolls 312, or a signal to make the rotation speed of the lower roll slower than the rotation speed of the upper roll.

In addition, when the difference between the increase in the first color ratio in the band-shaped area B52 and the decrease in the first color ratio in the band-shaped area A51 in the PLC9 is greater than the first critical value, the output is to make the rotation speed of the lower roll be higher than the upper The signal that the rotation speed of the rollers is slow, and when the difference is greater than the second critical value (larger than the first critical value), a signal to stop the pair of rollers 312 may also be output.

In addition, the PLC9 may output a warning signal with high urgency when the difference becomes larger.

As described above, according to the system of the present embodiment equipped with the band-shaped quantization processing unit 50, compared to the first embodiment, it is possible to more accurately support the operator in confirming the position or size of the machine and material. In addition, the PLC 9 uses the data quantized by the band-shaped quantization processing unit 50 to realize appropriate control according to the state of the subject. For example, changing the control value so that the machine as the subject can move to an appropriate position, or changing the control value to prevent the rolling material of the subject from colliding with sensors or machinery.

In addition, the system of Embodiment 2 described above can be combined with the system of Embodiment 1. This point is also the same in the following embodiments.

In addition, although the band-shaped quantization processing unit 50 of the second embodiment described above sets the band-shaped area in the horizontal direction, it may also set the band-shaped area in the vertical direction or the oblique direction.

In addition, the stripe quantization processing unit 50 of the second embodiment described above divides the screen into four, but it may also divide into two, three, or more than five.

Embodiment 3

Next, the third embodiment will be described with reference to Figs. 7 and 8. In the first embodiment described above, the ratio of the first color to the overall image is calculated. On the other hand, in Embodiment 3, the proportion of the first color in each grid-shaped area divided into a grid is calculated, and the change in the proportion of the first color in each grid-shaped area is compared in more detail. Analyze the state of the subject.

Fig. 7 is a block diagram of the image analysis device 1 of the third embodiment. The configuration shown in FIG. 7 is the same as that in FIG. 2 or FIG. 5 except that the image quantization unit 23 includes a grid-like quantization processing unit 60, so the description of the common configuration is omitted or simplified.

The image processing unit 22 binarizes the image viewed from the width direction of the material.

The grid-like quantization processing unit 60 calculates the proportion of the first color (for example, white) in each grid-like region obtained by dividing the image obtained by binarization by the image processing unit 22 into grids. In the example shown in Fig. 7, the binarized image is divided into 4 regions in the horizontal direction and 3 regions in the vertical direction, for a total of 12 regions. The divided areas are called grid-shaped area A61, grid-shaped area B62, grid-shaped area C63, grid-shaped area D64, grid-shaped area E65, grid-shaped area F66, grid-shaped area G67, grid-shaped area H68, grid-shaped area I69, The lattice-shaped area J70, the lattice-shaped area K71, and the lattice-shaped area L72. The grid-like quantization processing unit 60 calculates the ratio of the number of white pixels in each grid-like area.

Calculate the proportion of the first color in each grid area (hereinafter referred to as the first color ratio) by dividing the image into grids in real time. By confirming the change in the first color ratio, the position of the subject can be grasped in a quantitative way , Changes in shape.

For example, the rolled material 311 (refer to Fig. 3) used as a material may jump on the conveying roller 313 (refer to Fig. 3) of a steel plant. At this time, the imaging object (rolled material) moves in parallel from the grid-shaped area A61 to the grid-shaped area B62. Therefore, by confirming that the increase of the first color ratio of the grid-shaped area A61 is equal to the decrease of the first color ratio of the grid-shaped area B62, it can be confirmed that the photographic subject moves in parallel.

Moreover, when the photographic subject extends obliquely and extends from the grid-shaped area A61 to the grid-shaped area F66, the first color ratio of the grid-shaped area A61 is saturated, and the first color ratio of the grid-shaped area F66 increases. From the increase or decrease of the first color ratio of the adjacent grid-shaped area B62 and the increase or decrease of the first color ratio of the grid-shaped area E65, it can be confirmed whether the photographic subject extends without shaking horizontally.

In addition, the shape of the subject may be deformed. For example, in a hot rolling line and a thick plate line of a steel plant, the rolled material 311 (refer to Fig. 3) may be warped upward due to rolling. Refer to Figure 8 to explain an example of image analysis. t3 is the shape of the rolled material in the previous cycle, and t4 is the shape of the rolled material in the current cycle. The grid-like quantization processing unit 60 divides the binarized image into at least grid-like grid-like regions F66, grid-like regions J70 laterally adjacent to grid-like regions F66, and grid-like regions adjacent to above grid-like region F66 G67, the grid-shaped area K71 adjacent to the grid-shaped area J70 and adjacent to the grid-shaped area G67 in the lateral direction, and calculate the first color ratio of each area. The boundary between the grid-shaped area F66 and the grid-shaped area G67 is set above the rolled material.

The operator confirms the change of the first color ratio in each grid-shaped area accompanying the change from t3 to t4. By confirming that the first color ratio increases in the order of the grid-shaped area F66 and the grid-shaped area J70, and the first color ratio in the grid-shaped area K71 increases according to the decrease in the first color ratio in the grid-shaped area G67, it can be quantified Ways to control the upward warpage of the rolled material.

The PLC9 receives the data (the first color ratio in each grid-shaped area) output by the numerical output unit 24. PLC9 can perform control based on the position of the subject, control based on the amount of movement of the subject, and control based on the shape of the subject from the first color ratio in each grid-shaped area (e.g., control based on the amount of upward warpage) .

The control based on the amount of upward warpage will be described with reference to the above-mentioned Figure 8. PLC9 outputs a warning when the first color ratio increases in the order of grid area F66 and grid area J70, and the first color ratio in grid area K71 increases according to the decrease in the first color ratio in grid area G67 At least one of a signal and a control signal. The control signal changes the rotation speed of the pair of rolls 312 to suppress the upward warping of the rolled material. For example, the control signal is a signal to stop the pair of rolls 312, or a signal to make the rotation speed of the lower roll slower than the rotation speed of the upper roll.

As described above, according to the system of the present embodiment provided with the grid-like quantization processing unit 60, compared to the first embodiment, it is possible to more accurately support the operator in confirming the position or size of the machine and material. In addition, the PLC 9 uses the data quantized by the grid-like quantization processing unit 60 to realize appropriate control according to the state of the subject.

In addition, the system of Embodiment 3 described above can be combined with the system of Embodiment 1 and the system of Embodiment 2.

In addition, the grid-like quantization processing unit 60 of the third embodiment described above defines the regions in a grid-like shape in the horizontal and vertical directions, but the grid-like regions may be set in the oblique direction.

In addition, the grid-like quantization processing unit 60 of the third embodiment described above divides the screen into 12, and the number of divisions is not limited if the screen is equally divided. Especially when shooting a subject from an oblique angle, the image is converted into a portrait shot from the side or from above. At this time, the amount of movement and deformation of the subject may become smaller. In this case, by increasing the number of divisions in the grid pattern, the reduced amount of movement and deformation can be detected.

The embodiments of the present invention have been described above, but the present invention is not limited to the above-mentioned embodiments, and various modifications can be made without departing from the purpose of the present invention.

1‧‧‧Image analysis device

2‧‧‧The first surveillance camera

3‧‧‧Second surveillance camera

4‧‧‧Image data converter

5‧‧‧Signal line for portrait

6‧‧‧Network for dynamic portrait

7‧‧‧Data collection device

8‧‧‧Human Machine Interface (HMI)

9‧‧‧Programmable Logic Controller (PLC)

10‧‧‧Control network

21‧‧‧Dynamic Portrait Data Collection Department

22‧‧‧Image Processing Department

23‧‧‧Image Quantization Department

24‧‧‧Numerical output

Claims (4)

An image analysis device for industrial factories, which is equipped with: a dynamic image data collection section, which collects and shoots the dynamic image data of the machines that constitute the industrial plant and the materials processed in the machine in real time; The aforementioned dynamic portrait data extracts the portrait, converts the designated color to the first color, and converts colors other than the designated color to the second color, and then binarizes the portrait; the portrait quantization part is based on the conversion to the first color The number of pixels is used to quantify the binarized image; and the value output unit outputs the quantized data; the image quantization unit calculates that the binarized image is divided into parallel with the conveying direction of the material The proportion of the aforementioned first color in each band-shaped area. The image analysis device for an industrial factory described in the first item of the scope of patent application, wherein the image quantization unit calculates the first color in each grid area formed by dividing the binarized image into a grid proportion. A monitoring and control system for industrial factories, equipped with: image analysis device for industrial factories, equipped with: dynamic image data collection section, which collects and shoots the dynamic image data of the machines that constitute the industrial factory and the materials processed in the machine in real time; image processing Department, which extracts the portrait from the aforementioned dynamic portrait data every certain period, converts the designated color into the first color, and converts colors other than the designated color For the second color, the image is binarized; the image quantization unit quantifies the binarized image according to the number of pixels converted into the first color; and the numerical output unit outputs the quantized data ; And a programmable logic controller that controls the aforementioned machine; the aforementioned machine includes a pair of rolls that are rolled as the material of the aforementioned material; the aforementioned image processing section is to perform binary value of the aforementioned image of the aforementioned material from the width direction The image quantization unit divides the binarized image into at least the first area and the second area adjacent to the first area in parallel with the conveying direction of the rolling material, and calculates the foregoing in each area The proportion of the first color; the programmable logic controller is the increase in the proportion of the first color in the second area and the decrease in the proportion of the first color in the first area When the difference in the amount is greater than the critical value, at least one of a warning signal and a control signal is output. The control signal changes the rotation speed of the pair of rolls to suppress the upward warpage of the rolled material. A monitoring and control system for industrial factories, equipped with: image analysis device for industrial factories, equipped with: dynamic image data collection section, which collects and shoots the dynamic image data of the machines that constitute the industrial factory and the materials processed in the machine in real time; image processing Department, which extracts portraits from the aforementioned dynamic portrait data in every certain period, and refers to The fixed color is converted to the first color, and the color other than the specified color is converted to the second color, and the image is binarized; the image quantization unit quantifies the binary value based on the number of pixels converted into the first color And a numerical output unit that outputs the aforementioned quantified data; and a programmable logic controller that controls the aforementioned machine; the aforementioned machine includes a pair of rolls that are rolled as the aforementioned material; and the aforementioned image processing unit Binarizes the image of the rolling material viewed from the width direction; the image quantization unit divides the binarized image into at least a grid-like first area and a second area adjacent to the first area. Area, the third area adjacent to the above of the first area, and the fourth area adjacent to the above of the second area and adjacent to the horizontal direction of the third area, and calculate the proportion of the first color in each area ; The proportion of the programmable logic controller in the first color increases in the order of the first area and the second area, and the proportion of the first color in the fourth area is based on the first When the proportion of the first color in the zone 3 decreases and increases, at least one of a warning signal and a control signal is output. The control signal changes the rotation speed of the pair of rolls to suppress the upward warpage of the rolled material.

Images