Cold rolling on-line strip surface defect detection system

Abstract

The embodiment of the specification discloses a cold-rolling online strip steel surface defect detection system, an industrial area array camera is used for collecting the strip steel surface image of the target strip steel in the cold-rolling process, and the collected strip steel surface image is transmitted to a defect identification device; the defect identification device is used for identifying defects of the strip steel surface image after receiving the collected strip steel surface image, and the identified defect position is image intercepted to obtain a defect image, and the defect image is sent to a defect classification device; the defect classification device is used for processing the defect image by using a pre-trained defect classification model after obtaining the defect image, and obtaining the defect type corresponding to the defect image. The cold-rolling online strip steel surface defect detection system disclosed in the specification can improve the defect detection speed of the strip steel surface, so that the defect detection speed matches the production speed of the cold-rolling unit, and the defect detection efficiency is improved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present specification relate to the technical field of steel rolling, and particularly relate to a cold-rolling online strip surface defect detection system. BACKGROUND

[0002] With the improvement of rolling process level of metallurgical industry, the requirements for the surfaces of various types of strips are higher and higher, and the surface quality of the strip has become a direct performance index of the strip product.

[0003] In the prior art, the detection method of the surface of the cold-rolled strip is usually a manual visual detection method. Since the production speed of the existing cold-rolling unit is high, the detection speed of the manual visual detection method is slow, and the manual visual detection method is inefficient in detecting small surface defects and is easy to cause eye fatigue of the detection personnel, resulting in a high probability of detection errors. Therefore, a strip surface defect detection system that can match the production speed of the cold-rolling unit is urgently needed. SUMMARY

[0004] The cold-rolling online strip surface defect detection system provided by the embodiments of the present specification can improve the defect detection speed of the strip surface, match the defect detection speed with the production speed of the cold-rolling unit, and improve the defect detection efficiency.

[0005] The first aspect of the embodiments of the present specification provides a cold-rolling online strip surface defect detection system, comprising:

[0006] an industrial area array camera, configured to collect a strip surface image of a target strip in a cold-rolling process and transmit the collected strip surface image to a defect recognition device;

[0007] the defect recognition device, configured to perform defect recognition on the strip surface image after receiving the collected strip surface image, perform image cropping on the recognized defect position to obtain a defect image, and send the defect image to a defect classification device;

[0008] the defect classification device, configured to perform processing on the defect image by using a pre-trained defect classification model after obtaining the defect image, to obtain a defect type corresponding to the defect image.

[0009] Optionally, the defect recognition device is configured to perform defect recognition on the strip surface image by using a gray threshold segmentation method to recognize a defect position in the strip surface image; and perform image cropping on the strip surface image according to the defect position to obtain a defect image corresponding to the strip surface image.

[0010] Optionally, the defect classification device is configured to obtain the strip steel coil data corresponding to the strip steel surface image, the strip steel weld position, and the position of the defect on the strip steel after obtaining the defect image corresponding to the strip steel surface image.

[0011] Optionally, the defect classification device is configured to receive the strip steel weld position sent by a first communication device before obtaining the strip steel coil data corresponding to the strip steel surface image, the strip steel weld position, and the position of the defect on the strip steel, wherein the first communication device communicates with a controller arranged in the cold rolling production line and is configured to obtain the strip steel weld position of the target strip steel collected by the controller.

[0012] Optionally, the first communication device is configured to obtain the running speed of the cold rolling production line collected by the controller and transmit the running speed to the defect classification device.

[0013] Optionally, the first communication device communicates based on an OPC communication protocol.

[0014] Optionally, the defect classification device is configured to receive the strip steel coil information sent by a second communication device before obtaining the strip steel coil data corresponding to the strip steel surface image, the strip steel weld position, and the position of the defect on the strip steel, wherein the second communication device is configured to obtain the strip steel coil data of the target strip steel stored in the cold rolling production line, and the strip steel coil data includes strip steel raw material coil information.

[0015] Optionally, the second communication device is configured to obtain the finished product coil information of the target strip steel after cold rolling of the target strip steel is completed and transmit the finished product coil information to the defect classification device.

[0016] Optionally, the second communication device communicates based on a TCPIP communication protocol.

[0017] Optionally, the method further comprises:

[0018] The human-computer interaction device is configured to be connected with the defect classification device and display the defect image.

[0019] The beneficial effects of the embodiments of the present specification are as follows:

[0020] Based on the above technical scheme, the industrial area array camera is used to collect the strip surface image of the target strip steel in the cold rolling process, and the collected strip surface image is transmitted to the defect identification device; the defect identification device is used to identify the defects of the strip surface image, and the identified defect position is image intercepted to obtain a defect image, and the defect image is sent to the defect classification device; the defect classification device is used to process the defect image by using a pre-trained defect classification model to obtain the defect type corresponding to the defect image; in this way, the strip surface image is collected by the industrial area array camera, and the strip surface image is identified by using the defect identification device to obtain the defect image, and the defect image is processed by the defect classification model in the defect classification device to obtain the defect type corresponding to the defect image. The surface of the cold-rolled online strip steel is detected by the industrial area array camera, the defect identification device and the defect classification model, so that the surface of the cold-rolled online strip steel is automatically detected by the machine. Compared with the manual visual detection method of the prior art, the defect detection speed of the strip surface can be effectively improved, so that the defect detection speed matches the production speed of the cold rolling unit, and the defect detection efficiency is improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 The system architecture diagram of the cold-rolled online strip steel surface defect detection system in the embodiment of the present application is shown.

[0022] Figure 2 The connection structure diagram of each device in the cold-rolled online strip steel surface defect detection system in the embodiment of the present application is shown. DETAILED DESCRIPTION

[0023] In order to better understand the above technical scheme, the technical scheme of the embodiment of the present application will be described in detail below with reference to the drawings and specific embodiments. It should be understood that the specific features in the embodiment of the present application and the embodiments are detailed descriptions of the technical scheme of the present application, and are not limitations of the technical scheme of the present application. In the case of no conflict, the technical features in the embodiment of the present application and the embodiments can be combined with each other.

[0024] REFERENCE Figure 1 A cold-rolled online strip steel surface defect detection system of the present application is shown, which comprises:

[0025] The industrial area array camera 10 is used to collect the strip surface image of the target strip steel in the cold rolling process, and the collected strip surface image is transmitted to the defect identification device;

[0026] The defect identification device 20 is used to identify the defects of the strip surface image after receiving the collected strip surface image, and the identified defect position is image intercepted to obtain a defect image, and the defect image is sent to the defect classification device;

[0027] The defect classification device 30 is configured to, after obtaining the defect image, process the defect image by using the pre-trained defect classification model to obtain a defect type corresponding to the defect image.

[0028] In the embodiments of the present specification, the industrial area array camera 10 can be arranged above or beside the cold rolling production line to collect the strip surface image of the strip steel produced by the cold rolling production line in real time. Further, the target strip steel can be the strip steel currently produced by the cold rolling production line, or a pre-specified strip steel. When the target strip steel is the strip steel currently produced by the cold rolling production line, each strip steel produced by the cold rolling production line can be taken as the target strip steel, and the above operation is performed to identify the defect image corresponding to each strip steel and the defect type corresponding to the defect image.

[0029] After the cold rolling production line is started, the industrial area array camera 10 is started to collect the strip surface image of the target strip steel in the cold rolling process, and the collected strip surface image is transmitted to the defect identification device. At this time, according to the actual demand, the camera parameters of the industrial area array camera 10 also need to be set, including the exposure time, the image resolution, the trigger mode, etc. The exposure time of the industrial area array camera 10 can be 500us or 600us, etc. The image resolution column can be 1920x128px or 1680x128px, etc. The trigger mode is the rising edge external trigger, etc.

[0030] After the industrial area array camera 10 collects one strip surface image of the target strip steel, the collected strip surface image can be transmitted to the defect identification device 20, so that the defect identification device 20 can obtain the strip surface image of the target strip steel in real time. Of course, all the collected strip surface images can also be transmitted to the defect identification device every set time length, such as 1s or 2s, etc. At this time, each collected strip surface image needs to be sorted by time, so that the defect identification device 20 can obtain the strip surface image of the target strip steel with a time delay.

[0031] Further, after the defect identification device 20 obtains the strip surface image, the strip surface image is processed by using an image processing algorithm to identify the defect position in the strip surface image, and then the image in the strip surface image is intercepted according to the defect position to obtain a defect image corresponding to the strip surface image.

[0032] Specifically, when the defect identification device 20 processes the strip surface image by using the image processing algorithm, the strip surface image can be processed by using the gray threshold segmentation method to identify the defect position in the strip surface image. Then, the image in the strip surface image is intercepted according to the defect position to obtain a defect image corresponding to the strip surface image.

[0033] Of course, the image processing algorithm can further include a processing method such as Gaussian filtering after including the gray threshold segmentation method.

[0034] Specifically, the key parameters in the image processing algorithm include a Gaussian filtering frequency of 0.4 or 0.5, an image segmentation gray threshold of 20 or 18, a defect extraction gray threshold of 50 or 55, an extracted defect area threshold of 2 or 3, etc. First, the surface image of the strip steel is subjected to Gaussian filtering, and the frequency can be 0.4. Then, the surface image of the strip steel subjected to Gaussian filtering is processed by using the image segmentation method, and the image segmentation threshold can be 20. After the surface image of the strip steel processed by the image segmentation method is subjected to defect extraction, the gray threshold of the defect extraction can be 50. Then, image interception is performed based on the defect position of the defect extraction, and the area threshold of the image interception can be 2 or 3, etc., so as to obtain a defect image. Of course, the image size of the defect image can also be preset to a set value, which can be 100x100px or 200x200px, etc., so that the image size of the final obtained defect image is the set value.

[0035] In the embodiments of the present specification, the defect recognition device 20 can be provided with a GIGE interface, so that the defect recognition device 20 acquires the surface image of the strip steel collected by the industrial area array camera 10 in real time through the GIGE interface.

[0036] In addition, after the defect recognition device 20 recognizes the defect image, it sends the defect image recognized by it to the defect classification device 30.

[0037] Therefore, after receiving the defect image, the defect classification device 30 can also store the defect image in its own database, and of course, it can also be stored in the storage device connected to the defect classification device 30, so as to subsequently process the defect image. The following is a specific example of storing the defect image in the database.

[0038] In addition, before processing the defect image, the defect classification device 30 also needs to pre-train a defect classification model. The defect classification model can be an intelligent learning model trained based on a neural network algorithm. Specifically, the historical defect data of the surface of the strip steel can be used to train the model, and the model that meets the constraint condition is obtained by training as the defect classification model. In the middle, the input of the defect classification model is the defect image information, and the output result is the image type and name, and the data is marked in the database. In this way, the pre-trained defect classification model can realize accurate classification and labeling of the defect image.

[0039] In another embodiment, the defect classification device 30 can also be used to obtain the strip steel coil data, the strip steel weld position and the position of the defect on the strip steel corresponding to the strip steel surface image after obtaining the defect image corresponding to the strip steel surface image, and then store the defect image, the strip steel coil data, the strip steel weld position and the position of the defect on the strip steel in the database of the defect classification device 30 in association.

[0040] Specifically, the defect classification device 30 is configured to receive the strip steel weld position sent by a first communication device before obtaining the strip steel coil data, the strip steel weld position and the position of the defect on the strip steel corresponding to the strip steel surface image, wherein the first communication device communicates with a controller arranged in the cold rolling production line and is configured to obtain the strip steel weld position of the target strip steel collected by the controller.

[0041] The first communication device is configured to obtain the running speed of the cold rolling production line collected by the controller and transmit the running speed to the defect classification device 30.

[0042] Specifically, the first communication device can communicate based on the OPC communication protocol, communicate with the controller through the OPC communication protocol, obtain the running speed of the cold rolling production line and the strip steel weld position in real time, and transmit them to the defect classification device 30 and write them into the database of the defect classification device 30. Of course, the first communication device can also obtain the weld position conversion data of the target strip steel in real time and transmit it to the defect classification device 30 and then store it in the database. In this way, since the first communication device communicates based on the OPC communication protocol, efficient communication is achieved.

[0043] In addition, the defect classification device 30 can also be used to receive the strip steel coil information sent by a second communication device before obtaining the strip steel coil data, the strip steel weld position and the position of the defect on the strip steel corresponding to the strip steel surface image, wherein the second communication device is configured to obtain the strip steel coil data of the target strip steel stored in the cold rolling production line, and the strip steel coil data includes the raw material coil information of the strip steel.

[0044] The second communication device is configured to obtain the finished coil information of the target strip steel after the cold rolling of the target strip steel is completed, and transmit the finished coil information to the defect classification device.

[0045] Specifically, the second communication device can communicate based on the TCPIP communication protocol. In this way, the second communication device can communicate with the cold rolling production line through the TCPIP protocol and receive the raw material coil information and the finished coil information messages from the cold rolling production line; the raw material coil information is received when the steel coil welding is completed, and the finished coil information is received when the outlet steel coil is cut. In this way, since the second communication device communicates based on the TCPIP communication protocol, stable and accurate communication can be achieved.

[0046] In another embodiment, the cold-rolled on-line strip steel surface defect detection system can further comprise a human-computer interaction device connected with the defect classification device 30 for displaying the defect image.

[0047] Specifically, the human-computer interaction device contains a display screen, and of course, it can also contain a processor. The human-computer interaction device can acquire the defect image and its corresponding defect data stored in the database in real time, and display the defect image and its corresponding defect data in the form of a graphical interface and a table. The defect image and its image data can be refreshed in real time, including the location mark of the defect on the steel coil and the image corresponding to the defect. On the one hand, the human-computer interaction device can display the current coil defect image and its corresponding defect data, and on the other hand, it can also display the historical defect image and its corresponding defect data.

[0048] In actual application process, as shown in Figure 2 The connection structure diagram of each device in the cold-rolled on-line strip steel surface defect detection system provided by the embodiments of the present specification is shown. The industrial area array camera 10 is used to collect the strip steel surface image of the target strip steel in real time, and transmit the collected strip steel surface image to the defect recognition device 20 in real time. The defect recognition device 20 comprises a GIGE communication module 201, a defect recognition module 202 and an image sending module 203. The strip steel surface image sent by the industrial area array camera 10 is acquired in real time through the GIGE communication module 201. In this process, the parameters of the camera set in the module need to be set, i.e. the exposure time is 500us, the image size is 1920x128px, and the trigger mode is rising edge external trigger. Then, the strip steel surface image acquired in real time is processed at high speed using the image processing algorithm in the defect recognition module 202 to obtain the defect position information of the strip steel surface image. The key parameters in the image processing algorithm are as follows: the Gaussian filter frequency is 0.4, the image segmentation grayscale threshold is 20, the defect extraction grayscale threshold is 50, and the extracted defect area threshold is 2. Then, the defect is intercepted, the intercepted image size is 100x100px, the defect image is intercepted, and then the defect image is sent to the defect classification device 30 through the image sending module 203.

[0049] In addition, the defect classification device 30 comprises an image storage module 301, a defect classification module 302, an information processing module 303, a database 304, an L1 communication module 305 and an L2 communication module 306. The image storage module 301 is used to store the defect image sent by the image sending module 203 into the database 304, and the image storage module 301 is also used to match and store the information such as the steel coil corresponding to the defect image, the weld position, the position of the defect on the strip steel and the like into the database 304.

[0050] Further, the defect classification module 302 can establish an intelligent learning model through a neural network algorithm, train the model using the collected defect data, and obtain a defect classification model through training, wherein the input of the defect classification model is defect image information, the output result is image type and name, and the data is marked in the data.

[0051] Further, the information processing module 303 is used for automatically switching the steel coil information according to the weld position, archiving the strip steel information that has been produced, and matching the information of the raw material coil to the finished product coil for convenient viewing; the L1 communication module 305 is used for communicating with the production line PLC system (L1 system 50) through an OPC protocol, acquiring the speed and weld position data of the production line in real time, and writing the data into a database, and the L1 communication module 305 is also used for acquiring the conversion data of the weld position; the L2 communication module 306 is used for communicating with the production line L2 system 51 through a TCPIP protocol, and receiving the raw material coil information and the finished product coil information from the L2 system. When the steel coil is welded, the raw material coil information is received, and when the export steel coil is cut, the finished product coil information is received.

[0052] Further, the cold rolling on-line strip steel surface defect detection system further comprises a man-machine interactive device 40 connected with the defect classification device 30, which is used for displaying the defect information in a graphical interface and a table according to the real-time stored defect data information and image information. The defect information is real-time refreshed, including the position mark of the defect on the steel coil and the image corresponding to the defect. In addition, the man-machine interactive device 40 can display the current coil defect information and the historical steel coil defect information. The switching of the steel coil information is completed by the information processing module.

[0053] The beneficial effects of the embodiments of the present specification are as follows:

[0054] Based on the above technical scheme, the industrial area array camera is used to collect the strip surface image of the target strip steel in the cold rolling process, and the collected strip surface image is transmitted to the defect identification device; the defect identification device is used to identify the defects of the strip surface image, and the identified defect position is image intercepted to obtain a defect image, and the defect image is sent to the defect classification device; the defect classification device is used to process the defect image by using a pre-trained defect classification model to obtain the defect type corresponding to the defect image; in this way, the strip surface image is collected by the industrial area array camera, and then the strip surface image is identified by the defect identification device to obtain the defect image, and then the defect image is processed by the defect classification model in the defect classification device to obtain the defect type corresponding to the defect image, and the surface of the cold rolling online strip steel is detected by the industrial area array camera, the defect identification device and the defect classification model, so that the surface of the cold rolling online strip steel is automatically detected by the machine, compared with the manual visual detection method of the prior art, the defect detection speed of the strip surface can be effectively improved, the defect detection speed is matched with the production speed of the cold rolling unit, and the defect detection efficiency is improved.

[0055] Although preferred embodiments of the present specification have been described, those skilled in the art can make additional changes and modifications to the embodiments once they know the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present specification.

[0056] Obviously, those skilled in the art can make various modifications and variations to the present specification without departing from the spirit and scope of the present specification. Thus, if these modifications and variations of the present specification fall within the scope of the claims of the present specification and their equivalent technologies, the present specification also intends to include these modifications and variations.

Claims

1. A cold rolling on-line strip surface defect detection system, characterized by, The application relates to a cold-rolled steel strip surface defect identification method and device. The application comprises the following: an industrial area array camera is used for collecting a steel strip surface image of a target steel strip in a cold rolling process and transmitting the collected steel strip surface image to a defect identification device; the defect identification device is used for identifying defects in the steel strip surface image after receiving the collected steel strip surface image, performing image interception on the identified defect position to obtain a defect image, and sending the defect image to a defect classification device; the defect classification device is used for processing the defect image by using a pre-trained defect classification model after obtaining the defect image, so as to obtain a defect type corresponding to the defect image; the defect classification device is used for obtaining steel strip coil data, a steel strip weld position and a defect position on a steel strip corresponding to the steel strip surface image after obtaining the defect image corresponding to the steel strip surface image; the defect classification device is used for receiving the steel strip weld position sent by a first communication device before obtaining the steel strip surface image corresponding to the steel strip coil data, the steel strip weld position and the defect position on the steel strip, wherein the first communication device communicates with a controller arranged in a cold rolling production line and is used for obtaining the steel strip weld position of the target steel strip collected by the controller; the first communication device is used for obtaining a running speed of the cold rolling production line collected by the controller and transmitting the running speed to the defect classification device; the defect classification device is used for receiving the steel strip coil data sent by a second communication device before obtaining the steel strip surface image corresponding to the steel strip coil data, the steel strip weld position and the defect position on the steel strip, wherein the second communication device is used for obtaining the steel strip coil data of the target steel strip stored in the cold rolling production line, and the steel strip coil data comprises steel strip raw material coil information; 2. The system of claim 1, wherein, the second communication device is used for obtaining finished product coil information of the target steel strip after cold rolling of the target steel strip is completed and transmitting the finished product coil information to the defect classification device.

3. The system of claim 1, wherein, The defect identification device is used for identifying defects in the steel strip surface image by using a gray threshold segmentation method, identifying the defect position in the steel strip surface image, and performing image interception on the steel strip surface image according to the defect position to obtain a defect image corresponding to the steel strip surface image.

4. The system of claim 1, wherein, The first communication device communicates based on an OPC communication protocol.

5. The system of claim 4, wherein, The second communication device communicates based on a TCPIP communication protocol. The application further comprises the following: a man-machine interaction device is used for being connected with the defect classification device and is used for displaying the defect image.

Images