CN111429425B - Rolling mill entrance abnormity identification method based on deep learning - Google Patents
Rolling mill entrance abnormity identification method based on deep learning Download PDFInfo
- Publication number
- CN111429425B CN111429425B CN202010201714.4A CN202010201714A CN111429425B CN 111429425 B CN111429425 B CN 111429425B CN 202010201714 A CN202010201714 A CN 202010201714A CN 111429425 B CN111429425 B CN 111429425B
- Authority
- CN
- China
- Prior art keywords
- rolling mill
- deep learning
- billet
- steel
- inlet
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a rolling mill entrance steel biting abnormity identification method based on deep learning, which comprises the following steps: acquiring an abnormity identification model according to a billet sample image at the inlet of a rolling mill; inputting the real-time billet image of the rolling mill inlet into the abnormity identification model, acquiring the moving state of the billet at the rolling mill inlet, and acquiring abnormity information according to the moving state; the method can effectively avoid a series of problems caused by manual participation, accurately identify the abnormity and effectively ensure the quality of the steel billet.
Description
Technical Field
The invention relates to the field of ferrous metallurgy, in particular to a rolling mill inlet abnormity identification method based on deep learning.
Background
In the smelting process in the field of ferrous metallurgy, a steel rolling mill needs to be used for carrying out hot rolling and cold rolling treatment on steel, and a steel rolling operator needs to operate an input-output roller way of the steel rolling mill. In the hot rolling and cold rolling processes, the situation that the steel billet is not clamped and can not be stopped can occur, namely, the steel biting at the inlet of the rolling mill is abnormal, and once the abnormal steel biting at the inlet of the rolling mill occurs, the abnormal steel biting must be processed in time, so that the whole working efficiency is prevented from being influenced. However, most of the existing monitoring for the abnormal entrance of the rolling mill depends on manual work, so that the efficiency is low, and due to manual careless omission, part of abnormal conditions can not be processed in time easily, and the quality of steel billets is affected.
Disclosure of Invention
In view of the problems in the prior art, the invention provides a rolling mill inlet steel biting abnormity identification method based on deep learning, and mainly solves the problems that the rolling mill inlet abnormity identification depends on manual work and is low in identification efficiency.
In order to achieve the above and other objects, the present invention adopts the following technical solutions.
A rolling mill inlet steel biting abnormity identification method based on deep learning comprises the following steps:
acquiring an abnormity identification model according to a billet sample image at the inlet of a rolling mill;
and inputting the real-time billet image of the inlet of the rolling mill into the abnormity identification model, acquiring the moving state of the billet at the inlet of the rolling mill, and acquiring abnormity information according to the moving state.
Optionally, the position of the steel billet at the entrance of the rolling mill is detected in real time through the anomaly identification model, and the moving state of the steel billet is determined according to the positions of continuous multiple frames of real-time steel billet images.
Optionally, a position threshold value is set, and when the difference value of the tail positions of the steel billets is not larger than the position threshold value in the positions of the continuous multi-frame real-time steel billet images, the abnormal steel biting at the inlet of the rolling mill is judged, and early warning information is started.
Optionally, creating a training data set and a testing data set through the billet sample image;
using the training data set to train a deep learning neural network to obtain the neural network model;
and optimizing parameters of the neural network model through the test data set to obtain the abnormal recognition model.
Optionally, before performing model training, normalization processing is performed on the training data set and the sample image in the test data set, respectively.
Optionally, replacing the VGG-16 in the SDD network architecture with a MobileNetV2 network, and removing the last global average pooling layer, full connection layer, and Softmax layer of the MobileNetV2 network; and changing the last two fully-connected layers of the original VGG-16 into convolutional layers so as to construct the deep learning neural network.
Optionally, the learning rate of the deep learning neural network is set by an exponential decay method.
Optionally, after regularizing the weight of the deep learning neural network, introducing the regularized weight into a cost function of the original SSD network, and creating a new cost function.
Optionally, L2 norm regularization may be employed to create the new cost function, expressed as:
wherein, C 0 The cost function of the original SSD network is represented by w, weight and lambda is a coefficient of a regular term.
Optionally, the weights of the deep learning neural network are updated by using back propagation, and the update formula is as follows:
wherein, C 0 The method is a cost function of an original SSD network, w is weight, lambda is a coefficient of a regular term, and alpha and beta are constant coefficients.
As described above, the method for identifying the abnormal steel biting at the inlet of the rolling mill based on the deep learning has the following advantages.
The abnormity of the inlet of the rolling mill is automatically identified through an abnormity identification model, so that a series of problems caused by manual participation are avoided; the model identification can ensure the real-time and accuracy of identification, improve the exception handling efficiency and ensure the quality of steel billets.
Drawings
Fig. 1 is a flowchart of a method for identifying abnormal steel biting at an inlet of a rolling mill based on deep learning according to an embodiment of the present invention.
Fig. 2 is a schematic structural diagram of a network bottleneck layer of MobileNetV 2.
Fig. 3 is a schematic structural diagram of an SSD network framework.
Detailed Description
The following embodiments of the present invention are provided by way of specific examples, and other advantages and effects of the present invention will be readily apparent to those skilled in the art from the disclosure herein. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention, and the components related to the present invention are only shown in the drawings rather than drawn according to the number, shape and size of the components in actual implementation, and the type, quantity and proportion of the components in actual implementation may be changed freely, and the layout of the components may be more complicated.
Referring to fig. 1, the invention provides a method for identifying abnormal steel biting at an inlet of a rolling mill based on deep learning, which comprises the steps of S01-S02.
In step S01, an anomaly recognition model is obtained from the billet sample image at the mill entrance:
in one embodiment, an image acquisition module such as a camera arranged near the inlet of the rolling mill can be used for acquiring a billet sample image, and a billet in the acquired sample image is labeled so as to detect the position of the billet by adopting a supervised learning method according to the labeling information. Further, the annotated sample images may be entered into a database to create a training dataset and a testing dataset, respectively. Specifically, the sample image may be divided into a test data set and a training data set in a ratio of 1; the division ratio of the sample image can also be set according to actual requirements.
In one embodiment, the sample images in the training data set and the test data set may be normalized to normalize the gray-scale values of the pictures from 0 to 255 to 0 to 1, respectively. Specifically, the image normalization may adopt a maximum and minimum normalization method, a logarithmic function conversion method, or an inverse cotangent function conversion method, and the like, where the maximum and minimum normalization method is taken as an example, the following formula is:
where xi represents the image pixel point value and max (x) and min (x) represent the maximum and minimum values of the image pixel, respectively.
In an embodiment, the image quality of the normalized sample image may be enhanced, and specifically, the sample image may be cut, flipped, rotated, brightness adjusted, contrast adjusted, saturation adjusted, and the like, so as to ensure the sample image quality and improve the billet detection efficiency.
In one embodiment, the training data set processed by the steps is used for training a deep learning neural network to obtain a neural network model; and optimizing parameters of the neural network model through the test data set to obtain an optimal neural network model as an abnormal recognition model.
The deep learning neural network can adopt a MobileNet V2-SSD deep learning neural network. MobileNetV2 is an improvement over MobileNetV1, which proposes two new concepts: inverted Residual inversed Residual and Linear Bottleneck bottle. The Inverted Residual invoked Residual is mainly used to increase the extraction of image features to improve precision, and the Linear Bottleneck bottleeck is mainly used to avoid information loss of the nonlinear function ReLU. The core of MobileNetV2 consists of 17 bottlenecks, whose network structure is shown in table 1 (where t is the multiple of the ascending dimension inside the Bottleneck layer, c is the dimension of the output feature, n is the number of repetitions, s is the step size of the convolution, and k is the width scaling factor).
TABLE 1
See table 2 for details of the bottleneck layer. The input increases the dimension from the k dimension to the tk dimension through 1 × 1conv + ReLU layer, then down samples the image by 3 × 3conv + ReLU separable convolution (stride > 1) when the characteristic dimension is already the tk dimension, and finally decreases the dimension from tk to k' dimension by 1 × 1conv (no ReLU).
TABLE 2
Furthermore, as shown in fig. 2, for the bottleneck layer, when the convolution step size stride =1, the input is mapped into the output; when stride =2, no shortcut connects the input and output features. The ReLU _6 function, as an activation function, can be expressed as:
y＝relu6(x)＝min(max(x,0),6)
where x and y represent input and output, respectively.
SSD is a single-stage target detection algorithm, and targets with different frame sizes are predicted by using feature maps with different scales. The SSD network structure is divided into two parts: basic network + pyramid network, where the basic network is transformable. The underlying network of the original SSD is the top 4 network of the VGG-16, and the pyramid network is a simple convolutional network with gradually-reduced feature maps and consists of 5 parts. Please refer to fig. 3 for a specific network structure of the SSD.
In the embodiment, a MobileNet V2 network is adopted to replace VGG-16 in an SDD network architecture, and a final global average pooling layer, a full connection layer and a Softmax layer of the MobileNet V2 network are removed; and changing the last two fully-connected layers of the original VGG-16 into convolutional layers so as to construct the deep learning neural network. Specifically, by replacing VGG-16 in the original SSD network architecture with a MobileNetV2 network, the configuration from Conv0 to Conv13 is completely consistent with the MobileNetV2 model, except that the last global average pooling, full connection layer and Softmax layer of MobileNetV2 are removed, and FC6 and FC7 of the original VGG-16 are replaced with Conv6 and Conv7, respectively. The MobileNet V2-SSD deep learning neural network firstly uses the MobileNet V2 network to extract image characteristic output characteristic graphs, and then uses the SSD target detection algorithm to detect information on a plurality of characteristic graphs output by the MobileNet V2 network.
In one embodiment, to avoid overfitting, an exponential decay method may be used to set the learning rate of the deep learning neural network.
In an embodiment, the weights of the deep learning neural network may be introduced into the cost function of the original SSD network after being regularized, so as to create a new cost function. Specifically, L2 norm regularization can be employed to create a new cost function, represented as:
wherein, C 0 The cost function of the original SSD network is shown, w is weight, and lambda is a coefficient of a regular term.
In one embodiment, the weights of the deep learning neural network are updated by using back propagation, and the updating formula is as follows:
wherein, C 0 The method is a cost function of an original SSD network, w is weight, lambda is a coefficient of a regular term, and alpha and beta are constant coefficients.
Inputting a training data set into a built deep learning neural network, training a neural network model according to errors between a billet label marked in the training data set and a prediction frame obtained by the deep learning neural network, continuously converging a predicted value to the error direction of the label and the prediction frame when the training network passes multiple iterations, and then updating parameters into each layer according to a chain rule by back propagation. And each iteration reduces propagation errors as much as possible according to the optimization direction of gradient descent, and finally obtains the final target detection result of all billet images in the data set. Further, a neural network model with the highest accuracy is obtained as an abnormality recognition model by testing the accuracy of the target detection result on the data set.
In step S02, the real-time billet image at the entrance of the rolling mill is input into the anomaly identification model, the moving state of the billet at the entrance of the rolling mill is obtained, and the anomaly information is obtained according to the moving state.
In one embodiment, a single frame of billet image can be obtained from the real-time video stream as an input of the anomaly identification model, and the position coordinates of the billet at the inlet of the rolling mill in the single frame of image are detected through the anomaly identification model. Furthermore, through multiple times of identification, the change of the position coordinates of the steel billet in continuous multiframe steel billet images can be obtained, and further the moving state of the steel billet within a certain time range can be obtained. If the coordinate position of the billet can be detected through the abnormity identification model, the displacement of the inlet billet of the rolling mill moving in a period of time is obtained, and whether the inlet steel biting abnormity of the rolling mill occurs is judged. Normally, the billet at the inlet of the rolling mill moves forwards on the input and output roller ways of the fixed rolling mill, namely the billet at the inlet of the rolling mill has certain displacement within a period of time, and the color of the billet identification frame is set to be green. If the steel billet at the inlet of the rolling mill is stopped and does not move, the abnormal situation that the steel billet is bitten at the inlet of the rolling mill is indicated, at the moment, the color of the steel billet identification frame is set to be red, and whether the steel billet is bitten at the inlet of the rolling mill is judged through the coordinate position change of the steel billet at the inlet of the rolling mill of the front frame and the rear frame in the on-site real-time video identified by the model. Specifically, a position threshold value can be set, and when the difference value of the tail positions of the steel billets is not greater than the position threshold value in the positions of the continuous multi-frame real-time steel billet images, the abnormal steel biting at the inlet of the rolling mill is judged. For example, when the absolute value of the difference value of the horizontal coordinates of the tail parts of the rolling mill inlet billets which are separated by 80 frames in the real-time video of the site is not more than 2 percent of the length of the billets. Wherein the setting of the position threshold value can be adjusted according to the actual situation.
After the steel biting abnormity is identified, the early warning information is triggered and started, and related workers are informed to take corresponding measures in time.
The invention is operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
The invention may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
In conclusion, the method for identifying the steel biting abnormity at the inlet of the rolling mill based on deep learning realizes the identification of the steel biting abnormity at the inlet of the rolling mill under the industrial scene without human participation, the identification accuracy rate is more than 99%, the effect is excellent under the industrial scene of actual steel making, unprecedented leaps exist in the technical field of identification of the steel biting abnormity at the inlet of the rolling mill, and the production efficiency of a steel mill is improved. Therefore, the invention effectively overcomes various defects in the prior art and has high industrial utilization value.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Those skilled in the art can modify or change the above-described embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (8)
1. A rolling mill entrance steel biting abnormity identification method based on deep learning is characterized by comprising the following steps:
acquiring an abnormal recognition model according to a billet sample image at the inlet of a rolling mill;
inputting the real-time billet image of the rolling mill entrance into the abnormity identification model, and acquiring the moving state of the billet at the rolling mill entrance, wherein the method comprises the following steps: detecting the position of the steel billet at the inlet of the rolling mill in real time through the abnormal recognition model, determining the moving state of the steel billet according to the positions of continuous multi-frame real-time steel billet images, and acquiring abnormal information according to the moving state, wherein the abnormal information comprises the following steps: and setting a position threshold, judging that the steel biting at the inlet of the rolling mill is abnormal when the difference value of the tail positions of the steel billets is not greater than the position threshold in the positions of the continuous multi-frame real-time steel billet images, and starting early warning information.
2. The method for identifying the abnormal steel biting at the entrance of the rolling mill based on the deep learning of claim 1 is characterized in that a training data set and a testing data set are created through the billet sample image;
using the training data set to train a deep learning neural network to obtain the neural network model;
and optimizing parameters of the neural network model through the test data set to obtain the abnormal recognition model.
3. The method for identifying the abnormal steel biting at the entrance of the rolling mill based on the deep learning as claimed in claim 2, characterized in that before model training, normalization processing is respectively carried out on the training data set and the sample images in the test data set.
4. The deep learning-based method for identifying the abnormal steel biting at the inlet of the rolling mill is characterized in that a MobileNet V2 network is adopted to replace VGG-16 in an SDD network architecture, and a final global average pooling layer, a full connection layer and a Softmax layer of the MobileNet V2 network are removed; and changing the last two fully-connected layers of the original VGG-16 into convolutional layers so as to construct the deep learning neural network.
5. The rolling mill inlet steel biting abnormity identification method based on deep learning of claim 2, is characterized in that the learning rate of the deep learning neural network is set by an exponential decay method.
6. The rolling mill inlet steel biting abnormity identification method based on deep learning of claim 4 is characterized in that after the weight of the deep learning neural network is regularized, the weight is introduced into a cost function of an original SSD network to create a new cost function.
7. The method for identifying the abnormal steel biting at the entrance of the rolling mill based on the deep learning as claimed in claim 6, wherein the new cost function is created by adopting L2 norm regularization, and is expressed as:
wherein, C 0 The method is a cost function of an original SSD network, w is weight, lambda is a coefficient of a regular term, and n is repetition times.
8. The rolling mill inlet steel biting abnormity identification method based on deep learning of claim 6, is characterized in that the weights of the deep learning neural network are updated by adopting back propagation, and the update formula is as follows:
wherein, C 0 The method is a cost function of an original SSD network, w is weight, lambda is a coefficient of a regular term, alpha and beta are constant coefficients, and n is repetition number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010201714.4A CN111429425B (en) | 2020-03-20 | 2020-03-20 | Rolling mill entrance abnormity identification method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010201714.4A CN111429425B (en) | 2020-03-20 | 2020-03-20 | Rolling mill entrance abnormity identification method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111429425A CN111429425A (en) | 2020-07-17 |
| CN111429425B true CN111429425B (en) | 2023-04-07 |
Family
ID=71553469
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010201714.4A Active CN111429425B (en) | 2020-03-20 | 2020-03-20 | Rolling mill entrance abnormity identification method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111429425B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112037199A (en) * | 2020-08-31 | 2020-12-04 | 中冶赛迪重庆信息技术有限公司 | Hot rolled bar collecting and finishing roller way blanking detection method, system, medium and terminal |
| CN112053334B (en) * | 2020-08-31 | 2023-04-28 | 中冶赛迪信息技术(重庆)有限公司 | Double-steel identification method, system and medium for cold blank table of heating furnace |
| CN112037197A (en) * | 2020-08-31 | 2020-12-04 | 中冶赛迪重庆信息技术有限公司 | Hot-rolled bar cold-shearing material accumulation detection method, system and medium |
| CN112114571B (en) * | 2020-09-24 | 2021-11-30 | 中冶赛迪重庆信息技术有限公司 | Industrial data processing method, system and equipment |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2011050990A (en) * | 2009-09-02 | 2011-03-17 | Jfe Steel Corp | On-line diagnostic method by tracking sensor on hot rolling line |
| CN102836886A (en) * | 2011-06-24 | 2012-12-26 | 宝山钢铁股份有限公司 | Detection method of head and tail positions and moving direction of hot-rolled piece |
| CN103878184A (en) * | 2012-12-24 | 2014-06-25 | 上海梅山钢铁股份有限公司 | Hot rolling line material position remote tracking method |
| JP2018161668A (en) * | 2017-03-27 | 2018-10-18 | Ｊｆｅスチール株式会社 | Tracking device and tracking method of hot rolling material |
| CN108830903A (en) * | 2018-04-28 | 2018-11-16 | 杨晓春 | A kind of steel billet method for detecting position based on CNN |
| CN108856303A (en) * | 2017-05-16 | 2018-11-23 | 上海梅山钢铁股份有限公司 | Reliable hot-rolling finishing mill frame stings steel signal control method |
| CN110059558A (en) * | 2019-03-15 | 2019-07-26 | 江苏大学 | A kind of orchard barrier real-time detection method based on improvement SSD network |
| CN110280608A (en) * | 2019-07-19 | 2019-09-27 | 北京宇轩智能科技有限公司 | The intelligent accurate rolled piece tracking processing method of steel rolling |
| CN110280602A (en) * | 2019-07-19 | 2019-09-27 | 北京宇轩智能科技有限公司 | Intelligent steel rolling rolling assessment system and processing method extremely |
| CN110517317A (en) * | 2019-08-28 | 2019-11-29 | 武汉西尔塔信息技术有限公司 | A kind of steel billet localization method and application based on the processing of machine vision camera image |
-
2020
- 2020-03-20 CN CN202010201714.4A patent/CN111429425B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2011050990A (en) * | 2009-09-02 | 2011-03-17 | Jfe Steel Corp | On-line diagnostic method by tracking sensor on hot rolling line |
| CN102836886A (en) * | 2011-06-24 | 2012-12-26 | 宝山钢铁股份有限公司 | Detection method of head and tail positions and moving direction of hot-rolled piece |
| CN103878184A (en) * | 2012-12-24 | 2014-06-25 | 上海梅山钢铁股份有限公司 | Hot rolling line material position remote tracking method |
| JP2018161668A (en) * | 2017-03-27 | 2018-10-18 | Ｊｆｅスチール株式会社 | Tracking device and tracking method of hot rolling material |
| CN108856303A (en) * | 2017-05-16 | 2018-11-23 | 上海梅山钢铁股份有限公司 | Reliable hot-rolling finishing mill frame stings steel signal control method |
| CN108830903A (en) * | 2018-04-28 | 2018-11-16 | 杨晓春 | A kind of steel billet method for detecting position based on CNN |
| CN110059558A (en) * | 2019-03-15 | 2019-07-26 | 江苏大学 | A kind of orchard barrier real-time detection method based on improvement SSD network |
| CN110280608A (en) * | 2019-07-19 | 2019-09-27 | 北京宇轩智能科技有限公司 | The intelligent accurate rolled piece tracking processing method of steel rolling |
| CN110280602A (en) * | 2019-07-19 | 2019-09-27 | 北京宇轩智能科技有限公司 | Intelligent steel rolling rolling assessment system and processing method extremely |
| CN110517317A (en) * | 2019-08-28 | 2019-11-29 | 武汉西尔塔信息技术有限公司 | A kind of steel billet localization method and application based on the processing of machine vision camera image |
Non-Patent Citations (5)
| Title |
|---|
| 620mm热连轧窄带钢生产线快节奏轧制的全线跟踪;曹剑钊等;《中国冶金》;20131215;第23卷(第12期);第29-42页 * |
| Fault detection and isolation of bearings in a drive reducer of a hot steel rolling mill;Marcello Farina 等;《Control Engineering Practice》;20150302;第35-44页 * |
| 物料跟踪在棒材生产中的实现;赵程;《柳钢科技》;20130815(第04期);第14-16页 * |
| 线、棒材轧机主轧线物料跟踪功能分析;李松辽等;《冶金自动化》;20010225(第02期);第24-28页 * |
| 轧钢过程仿真和跟踪系统的设计;王维;《工业计量》;20121126;第22卷(第06期);第18-21+25页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111429425A (en) | 2020-07-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111429425B (en) | Rolling mill entrance abnormity identification method based on deep learning | |
| CN111429424B (en) | Heating furnace inlet anomaly identification method based on deep learning | |
| CN112465748A (en) | Neural network based crack identification method, device, equipment and storage medium | |
| CN111524112B (en) | Steel chasing identification method, system, equipment and medium | |
| Li et al. | A multi-scale cucumber disease detection method in natural scenes based on YOLOv5 | |
| CN111860439A (en) | Unmanned aerial vehicle inspection image defect detection method, system and equipment | |
| CN114581782B (en) | Fine defect detection method based on coarse-to-fine detection strategy | |
| CN115331172A (en) | Workshop dangerous behavior recognition alarm method and system based on monitoring video | |
| CN111310671A (en) | Heating furnace bottom sump abnormity identification method, system and equipment based on deep learning | |
| CN115994900A (en) | Unsupervised defect detection method and system based on transfer learning and storage medium | |
| CN116152254A (en) | Industrial leakage target gas detection model training method, detection method and electronic equipment | |
| CN112308825A (en) | SqueezeNet-based crop leaf disease identification method | |
| CN111724338B (en) | Turntable abnormity identification method, system, electronic equipment and medium | |
| Mirani et al. | Object Recognition in Different Lighting Conditions at Various Angles by Deep Learning Method | |
| CN112766305B (en) | Visual SLAM closed loop detection method based on end-to-end measurement network | |
| CN113221812A (en) | Training method of face key point detection model and face key point detection method | |
| CN117011640A (en) | Model distillation real-time target detection method and device based on pseudo tag filtering | |
| CN111047654A (en) | High-definition high-speed video background modeling method based on color information | |
| Huang et al. | The detection of defects in ceramic cell phone backplane with embedded system | |
| CN115908843A (en) | Superheat degree recognition model training method, recognition method, equipment and storage medium | |
| CN114359300B (en) | Optimization method, device and system of image segmentation model and storage medium | |
| CN115376195A (en) | Method for training multi-scale network model and method for detecting key points of human face | |
| Mazni et al. | An investigation into real-time surface crack classification and measurement for structural health monitoring using transfer learning convolutional neural networks and Otsu method | |
| CN114998222A (en) | Automobile differential shell surface detection method, electronic equipment and medium | |
| CN111724337A (en) | Cold bed top punching identification method and system, electronic equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 401329 No. 5-6, building 2, No. 66, Nongke Avenue, Baishiyi Town, Jiulongpo District, Chongqing Applicant after: MCC CCID information technology (Chongqing) Co.,Ltd. Address before: Building 1, No. 11, Huijin Road, North New District, Yubei District, Chongqing Applicant before: CISDI CHONGQING INFORMATION TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |