A scrap steel briquette information collection system and identification method

Abstract

The present invention discloses a scrap steel briquette information acquisition system and an identification method, belonging to the technical field of scrap steel briquette quality control. The system of the present invention includes an image acquisition device, a rotating mechanism, a flipping mechanism, a weighing device and a monitoring device. The scrap steel briquette is placed on the flipping mechanism, and a weighing device is provided below the flipping mechanism. The weighing device together with the flipping mechanism is integrally installed on the rotating mechanism, and the rotating mechanism drives the scrap steel briquette to rotate; the image acquisition device is arranged on one side of the scrap steel briquette for acquiring the image information of the side of the scrap steel briquette; the monitoring device is used to monitor the process of packing the scrap steel briquette and acquiring the image information to prevent adulteration, and the identification method of the present invention uses a convolutional neural network model improved by transfer learning to identify the identity of the target scrap steel briquette, and can identify the identity information of the scrap steel briquette in real time and accurately to realize the intelligent identification of the composition of the scrap steel briquette.

Description

Technical Field

[0001] The present invention belongs to the technical field of scrap steel briquette quality control, and more specifically, relates to a scrap steel briquette information collection system and an identification method. Background Art

[0002] Scrap steel is a kind of recycled resource with energy conservation, low carbon environmental protection and recyclability. It is regarded as the only important raw material that can replace iron ore for steelmaking, and is also a very important strategic resource, which has been widely used in steel production. According to the data released by the China Iron and Steel Association, the crude steel output in China reached 928 million tons in 2018, setting a record high. According to the statistics of the China Scrap Steel Application Association, the generation amount of scrap steel resources in China was about 220 million tons in 2018. The World Steel Association predicts that the generation amount of scrap steel resources in China will reach 300 million tons in 2030 and 400 million tons in 2050. In the future, the scrap steel resources in China may not only supply the production and manufacturing of domestic steel enterprises, but also become an essential important raw material for the development of the world steel industry. There is no doubt that China ranks first in the world in terms of both crude steel output and scrap steel output, and will still maintain a rapid growth trend in the long term.

[0003] In order to facilitate the use in steel mills, some treatments are generally made to scrap steel in advance, such as forming scrap steel briquettes by extruding scrap steel through a baling machine. Compared with directly using scrap steel, using scrap steel briquettes can improve the efficiency of steel mills in storage and loading and unloading, and at the same time obtain high-quality molten iron with low impurity elements, so as to produce high-grade products. However, there is also a quality problem with scrap steel briquettes. Since there are many processing enterprises of scrap steel briquettes, some enterprises have the phenomenon of passing off inferior goods as good ones. Even some enterprises adopt the method of doping and forging in the central part of the scrap steel briquette, disturbing the supply market of briquettes and causing many steel mills to suffer losses, which greatly limits the wide use of scrap steel briquettes.

[0004] After retrieval, there are relatively few patents on scrap steel briquettes at present, mainly focusing on scrap steel type identification and scrap steel baling machine equipment. For example, the patent documents with Chinese patent publication numbers CN217375190U, CN217532032U and CN212599320U. Among these three patents, the first two patents can effectively save the packing time of scrap steel briquettes and improve the packing efficiency of scrap steel briquettes by carrying out structural design and optimization on the scrap steel packing device. The third patent mainly optimizes the inspection equipment for scrap steel briquettes, improving the shearing quality and shearing efficiency of the inspection equipment. However, it ultimately relies on manual experience to identify whether there is doping and forging in the scrap steel briquette, and it is difficult to improve the identification efficiency. Therefore, there is currently no good method for effectively supervising whether there is doping and forging in the scrap steel briquette packing process and for intelligently and automatically identifying the identity of the scrap steel briquette. Summary of the Invention

[0005] 1. Problem to be Solved

[0006] In view of the deficiencies proposed in the above-mentioned background art, the present invention provides a scrap steel briquette information acquisition system. Using this system, not only can the automated, intelligent, and high-efficiency acquisition of scrap steel briquette image information be realized, but also the doping and fraud during the scrap steel briquette packaging process can be supervised in real time;

[0007] Meanwhile, the present invention also provides a method for identifying scrap steel briquette information. By analyzing and processing the collected scrap steel briquette image information, the types of scrap steel briquettes can be quickly and accurately identified, thereby realizing the intelligent identification of the identity of scrap steel briquettes and facilitating the quality traceability of scrap steel briquettes.

[0008] 2. Technical Solution

[0009] To solve the above problems, the technical solutions adopted by the present invention are as follows:

[0010] A scrap steel briquette information acquisition system provided by the present invention includes an image acquisition device, a rotating mechanism, a flipping mechanism, a weighing device, and a monitoring device. The scrap steel briquette is placed on the flipping mechanism, and a weighing device is provided below the flipping mechanism. The weighing device together with the flipping mechanism is integrally installed on the rotating mechanism, and the rotating mechanism drives the scrap steel briquette to rotate; the image acquisition device is arranged on one side of the scrap steel briquette for acquiring the image information of the side of the scrap steel briquette; the monitoring device is used to monitor the scrap steel briquette packaging and image information acquisition processes to prevent doping in the scrap steel briquette during the packaging process. At the same time, by setting the image acquisition device, the surface image information of the scrap steel briquette can be automatically acquired, which is convenient for the automated determination of the types of scrap steel and the analysis of their component compositions, enabling the steel mill to use scrap steel briquettes more fully and reasonably in the production process, saving the production costs of the enterprise, and achieving the maximization of benefits.

[0011] As a further improvement of the present invention, the scrap steel briquette information acquisition system of the present invention further includes a control computer and a cloud server. The image acquisition device, rotating mechanism, flipping mechanism, weighing device, and monitoring device are all connected to the control computer through a controller, realizing automated operation. And the control computer is connected to the cloud server, and the cloud server is connected to a mobile device, so that the relevant data generated during the scrap steel briquette packaging, image acquisition, etc. can be transmitted to the cloud and can be viewed in real time through the terminal mobile device, further improving the convenience of using scrap steel briquettes.

[0012] As a further improvement of the present invention, the rotating mechanism includes a driving motor, a gear, and a rotating column. The driving motor is drivingly connected to one end of the rotating column through the gear, and the other end of the rotating column is fixedly connected to the weighing device. The driving motor and the weighing device are connected to the control computer. Specifically, a first gear is fixedly installed at the output end of the driving motor, a second gear meshing with the first gear is installed on the rotating column, a weighing device is fixedly installed at the top of the rotating column, and the weighing device is fixedly connected to the flipping mechanism. The rotating mechanism can be controlled to work through the control computer. During operation, the first gear is driven to rotate by the driving motor inside, the first gear then drives the second gear, and then the rotation of the second gear drives the rotating column to rotate, so that the rotating column drives the weighing device, the flipping mechanism, and the scrap steel briquette above to rotate together.

[0013] The weighing device is arranged at the middle position between the rotating column and the flipping mechanism. The weighing device can be controlled to work through the control computer. During operation, it will measure the weight data of the scrap steel briquette and then upload it to the control computer.

[0014] The flipping mechanism includes a mounting shaft and two symmetrically arranged hydraulic telescopic rods. The movable ends of the two hydraulic telescopic rods are respectively fixedly connected to a flap. The flipping angle range of each flap is 0° to 90°, and both flaps are rotatably installed on the mounting shaft. The hydraulic telescopic rods are connected to the control computer through a controller. The flipping mechanism can be controlled to work through the control computer. During operation, the flipping function of the scrap steel briquette is realized by controlling the two flaps to flip around the mounting shaft, and it rotates in cooperation with the rotating mechanism, which is convenient for the image acquisition device to collect the image information of the six faces of the scrap steel briquette and transmit the collected image data to the control computer. The control computer has the functions of image acquisition and processing, and can control the video whole-process monitoring device and the image acquisition device, so as to realize the image acquisition of the scrap steel briquette and the video recording function of the whole working process.

[0015] As a further improvement of the present invention, the rotating mechanism can achieve a full 360° rotation under the control of the control computer, so as to drive the weighing device, the flipping mechanism, and the scrap steel briquette arranged above the rotating mechanism to rotate together, so as to facilitate the image acquisition device for the scrap steel briquette to collect the image information of each face around.

[0016] As a further improvement of the present invention, the weighing device can work under the control of the control computer. Before the scrap steel briquette is placed, the control computer controls the weighing device to perform an automatic zeroing operation. Then, after the scrap steel briquette is placed, the weighing device will collect the weight of the scrap steel briquette and automatically upload it to the control computer.

[0017] As a further improvement of the present invention, the flipping mechanism can work under the control of a control computer. By controlling the extension and retraction of two hydraulic telescopic rods, the two flap plates are flipped and lifted around the mounting shaft. The maximum lifting angle of each flap plate is 90°, that is, perpendicular to the horizontal plane. At the same time, it should be noted that during the use of the two flap plates, the included angle formed between the two flap plates cannot be less than 90° to prevent squeezing the scrap steel briquette and causing damage to the equipment and products. Specifically, there are two working modes between the two flap plates:

[0018] "Left" mode: Place the scrap steel briquette on the right flap plate, start the rotating mechanism to collect images of the four vertical sides of the scrap steel briquette, and then control the right flap plate and the left flap plate to flip and lift upward, and the two flap plates gradually approach each other. When both the right flap plate and the left flap plate are lifted by 45°, the two flap plates can just clamp the scrap steel briquette, and then continue to slowly lift the right flap plate while the left flap plate descends until the right flap plate is lifted by 90° and the left flap plate is completely laid flat, and then start to lower the right flap plate until the right flap plate is completely laid flat, then the flipping of the scrap steel briquette can be completed. The bottom and top of the initial position of the scrap steel briquette are also flipped into vertical sides, and in cooperation with the rotation of the rotating mechanism, the image information of the six sides of the scrap steel briquette can be collected.

[0019] The flipping process of the "right" mode is the same as that of the "left" mode. That is, according to which side flap plate the image acquisition device is placed on, the corresponding mode is selected for image acquisition. For example, when the image acquisition device is placed close to the right flap plate side, the "left" mode is used for work, and vice versa, the "right" mode is used for work.

[0020] As a further improvement of the present invention, the monitoring device can work under the control of a control computer, and monitor the whole process of scrap steel briquette packing, flipping, weight measurement and image acquisition in real time, and then upload the whole process video of each scrap steel briquette to the control computer for storage separately.

[0021] As a further improvement of the present invention, the image acquisition device can work under the control of a control computer, cooperate with the rotating mechanism and the flipping mechanism to complete the acquisition of the image information of the six sides of the scrap steel briquette, and upload the collected image data to the control computer. The image acquisition device is installed and fixed through a bracket. More optimally, the bracket in the present invention is a telescopic bracket, which is convenient for adjusting the installation height of the image acquisition device so as to better acquire images of scrap steel briquettes of different sizes.

[0022] As a further improvement of the present invention, after obtaining the image data of the scrap steel briquette, the control computer can perform image processing on it, make predictions through a convolutional neural network model improved by transfer learning, analyze the types of scrap steel that make up the briquette, and then record the type information on the control computer. In the present invention, an algorithm program dedicated to scrap steel briquettes is installed in the control computer. This algorithm model is based on the convolutional neural network in deep learning, and then the parameter-transfer technology in transfer learning is used to improve it. The improved algorithm model can fully meet the requirement of accurately identifying the identity of scrap steel briquettes. Therefore, compared with the existing conventional image recognition algorithms, it not only significantly improves the accuracy of scrap steel recognition, but also improves the recognition speed.

[0023] Specifically, the algorithm model adopted in the present invention has been comprehensively improved for the professional data set of scrap steel briquette recognition, and the currently most popular transfer learning algorithm is incorporated into it, making the recognition of scrap steel briquettes by this algorithm model fast and accurate. Combined with Figure 4 , the specific improvements of the algorithm model of the present invention include the following processes:

[0024] 1) First, perform image transformation and data augmentation on the scrap steel briquette data set, and then divide the training set, validation set, and test set according to the ratio of 6:2:2;

[0025] 2) Delete the top classification structure of the convolutional neural network, then first use the Flatten layer to flatten the data of the multi-dimensional array into the data of a one-dimensional array, and then use the Linear layer to connect to the required number of classes. Note that the softmax layer is not needed at the end because the nn.CrossEntropyLoss function defaults to calculating the softmax layer inside the loss, as shown in Equation (1);

[0026]

[0027] In Equation (1): x is a vector of num_class dimensions; class is the category of the sample; num_class refers to the number of categories; x[class] is the predicted probability of the correct category, and x[j] is the predicted probability of each category. The cross-entropy loss function only calculates the loss of the correct category.

[0028] 3) Transfer the network weight parameters pre-trained on the ImageNet dataset to the prepared algorithm model, then freeze the network parameters before the Flatten layer and the Linear layer, and only train the weight parameters of other layers. The Adam algorithm is used for parameter optimization, which combines the advantages of the AdaGrad and RMSProp optimization algorithms, and dynamically adjusts the learning rate of each parameter according to the first-order moment estimate and second-order moment estimate of the gradient according to Equations (2 - 7);

[0029]

[0030] m t = β1m t-1 + (1 - β1)g t (3)

[0031]

[0032] In Equations (2) - (7), t is the time step, initialized to 0; θ is the parameter to be updated; f(θ) is the stochastic objective function of the parameter θ; g t is the gradient at time step t; β1 and β2 are the exponential decay rates of the first-order moment and second-order moment respectively; m t and v t are the first-order moment estimate and second-order moment estimate of the gradient respectively; and are the corrections for m t and v t respectively; ε is a constant added to maintain numerical stability; η is the set learning rate.

[0033] 4) Calculate the accuracy of the validation set once per epoch of training, save the model with the highest validation accuracy. After the specified number of epochs are completed, use the model with the highest validation accuracy to calculate the data of the test set to obtain the final test accuracy.

[0034] As a further improvement of the present invention, the cloud server has the ability to store, process and transmit massive data. After the control computer processes and summarizes all the information about the scrap steel briquettes collected from the weighing device, monitoring device and image acquisition device, it will send an application for new scrap steel briquette information to the cloud server. The cloud server will automatically review whether the scrap steel information in the application meets the requirements of the specified data structure. If it meets the requirements, it will accept the application from the control computer and store the information of the new briquette in the database of the cloud server; if it does not meet the requirements, it will return the requirement to the control computer to modify the information data structure, so as to realize the cloud storage of the scrap steel briquette data information.

[0035] As a further improvement of the present invention, after being authorized, the mobile device has the permission to access the cloud server at any time. When there is a need to identify scrap steel compacts, the user can also use the mobile device to take pictures of the target compact. Then, the mobile device processes the collected images and uploads them to the cloud server. The cloud server then predicts the target scrap steel compact through the convolutional neural network model improved by transfer learning, gives the information of the target scrap steel compact, and then displays all the relevant information of the target scrap steel compact on the mobile device for the user to trace the quality of the target scrap steel compact.

[0036] A method for identifying scrap steel compact information according to the present invention includes the following steps:

[0037] Step A: The monitoring device is turned on, and the scrap steel is packed by a packing machine to obtain a scrap steel compact;

[0038] Step B: The scrap steel compact is placed on the flipping mechanism. According to the setting of the control computer, the rotating mechanism and the flipping mechanism are automatically driven to cooperate so that the image acquisition device can smoothly acquire the images of six faces of the scrap steel compact. At the same time, the weighing device measures the weight of the scrap steel compact and transmits the acquired image information and weight data to the control computer;

[0039] Specifically, the method for placing the scrap steel is as follows:

[0040] First, lift the left and right turning plates simultaneously to an angle of 45° with the horizontal plane, place the scrap steel compact between the two turning plates, then lower one of the turning plates to the horizontal, and continue to lift the other turning plate to the vertical state. Finally, lower and flatten the vertical turning plate to complete the placement of the scrap steel compact.

[0041] The method for collecting six faces is as follows:

[0042] After the placement of the scrap steel compact is completed, start the rotating mechanism. Each time it rotates 90° clockwise (or counterclockwise), the image acquisition device acquires a side image of the scrap steel. After the scrap steel compact rotates one week, the image information of four sides is acquired. Then, start the flipping mechanism to work, and select the "left" mode or "right" mode for flipping according to the placement position of the image acquisition device to complete the acquisition of the image information of the bottom and top of the scrap steel compact.

[0043] Step C: The monitoring device stores the video of the whole process of packing, flipping, weight measuring and image acquisition of the photographed scrap steel compact into the control computer;

[0044] Step D: Remove the current scrap steel briquette, and place another scrap steel briquette to be information - collected stably on the turning plate. Repeat the above operations. Meanwhile, all the collected relevant data can be analyzed and processed by the control computer, and the quality judgment result can be output. The operator can also take pictures of the scrap steel briquette through a mobile device and then upload it to the cloud server. Through the convolutional neural network model improved by transfer learning, it can be predicted, and the information of the target scrap steel briquette can be given, so that the user can trace the quality of the target scrap steel briquette. The user uploads the picture taken by the mobile phone to the cloud, and it is transmitted to the control computer by the cloud for processing, which can quickly and accurately identify the identity information of the scrap steel, including the type and weight of the scrap steel and store it. At the same time, the relevant information of the scrap steel briquette can be incorporated into the database of the steel mill's intelligent batching system, and the scrap steel briquettes matching the smelting requirements in the database can be retrieved through the existing search algorithm for intelligent batching.

[0045] 3. Beneficial Effects

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0047] (1) For the scrap steel briquette information acquisition system of the present invention, through the optimized design of the system composition, especially through the integration of the rotating mechanism, flipping mechanism, weighing device and image acquisition device, it can not only automatically flip and rotate, facilitating the rapid acquisition of the image information of 6 faces of the scrap steel briquette, but also complete the automatic weighing during the process of image information acquisition, thus significantly improving the automation degree and efficiency of data acquisition;

[0048] (2) For the scrap steel briquette information acquisition system of the present invention, when the end - user steel mill receives the scrap steel briquette, it can identify the scrap steel briquette through a mobile device, query the video, weight and predicted scrap steel type during the processing of the briquette, and judge the quality of the scrap steel briquette, which can effectively reduce the judgment cost of the steel mill for the quality of the scrap steel briquette and is also convenient for controlling the overall supply - chain quality of the scrap steel briquette;

[0049] (3) For the scrap steel briquette information recognition method of the present invention, according to the characteristics that the appearances of different types of scrap steel briquettes are different, the convolutional neural network model improved by transfer learning is used to classify and recognize the scrap steel briquettes, and the type of scrap steel in the scrap steel briquette can be accurately analyzed. When batching before smelting in the steel mill, according to the differences in the steel types to be smelted and the sizes of the steelmaking furnaces used, the briquettes with the most suitable composition and weight in the database are selected for intelligent batching, and the corresponding briquettes are found through querying the database for intelligent batching, giving full play to the characteristics of the scrap steel briquettes, which can effectively reduce the smelting cost of the steel mill. Description of the Drawings

[0050] Figure 1Schematic diagram of the overall structure of an information collection system for scrap steel compacts according to the present invention;

[0051] Figure 2 Schematic diagram of the structure of the flipping mechanism in the present invention;

[0052] Figure 3 Schematic diagram of the structure of the rotating mechanism in the present invention;

[0053] Figure 4 Flowchart of the recognition of scrap steel compacts by the algorithm model of the present invention;

[0054] Figure 5 Six - face image information of the scrap steel compact sample 1 collected in Example 1 of the present invention;

[0055] Figure 6 Six - face image information of the scrap steel compact sample 2 collected in Example 2 of the present invention;

[0056] Figure 7 Six - face image information of the scrap steel compact sample 3 collected in Example 3 of the present invention;

[0057] In the figure:

[0058] 1. Control computer; 2. Image acquisition device; 3. Bracket;

[0059] 4. Rotating mechanism; 401. Driving motor; 402. First gear; 403. Second gear; 404. Rotating column;

[0060] 5. Flipping mechanism; 501. First hydraulic telescopic rod; 502. Second hydraulic telescopic rod; 503. First flap; 504. Second flap;

[0061] 6. Weighing device; 7. Mounting shaft; 8. Scrap steel compact; 9. Monitoring device; 10. Cloud server; 11. Mobile device. Detailed implementation manners

[0062] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Therefore, the detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the protection scope of the present invention.

[0063] As Figure 1As shown in the figure, the scrap steel briquette information acquisition system of the present invention includes a control computer 1, an image acquisition device 2, a rotating mechanism 4, a flipping mechanism 5, a weighing device 6, a monitoring device 9, a cloud server 10, and a mobile device 11. Among them, the control computer 1 is connected to the image acquisition device 2, the rotating mechanism 4, the flipping mechanism 5, and the weighing device 6 through a controller, facilitating the automatic acquisition of the image information of the scrap steel briquette 8. At the same time, the control computer 1 is also connected to the monitoring device 9 and the cloud server 10, capable of real-time monitoring whether there is adulteration behavior during the process of packing the scrap steel briquette and acquiring image information, playing an effective supervision role. At the same time, it can also upload relevant video data to the cloud server 10, and users can view the relevant information of the scrap steel briquette 8 at any time through the mobile device 11, improving the transparency of the scrap steel briquette production and preventing adulteration.

[0064] The above control computer 1, image acquisition device 2, weighing device 6, monitoring device 9, cloud server 10, and mobile device 11 are of existing types and can be directly purchased on the market. Such as Figure 2 and Figure 3 As shown in the figure, the rotating mechanism 4 in the present invention includes a housing, and a rotating column 404 is rotatably installed inside the housing. A second gear 403 is fixedly installed on the circumferential side wall of the lower part of the rotating column 404, and the top of the rotating column 404 passes through the housing of the rotating mechanism 4 and is fixedly connected to the bottom of the weighing device 6. A driving motor 401 is also fixedly installed inside the housing of the rotating mechanism 4. The driving end of the driving motor 401 is fixedly connected to a first gear 402, and the first gear 402 meshes with the second gear 403. By the rotation of the motor 401, the two gears are driven to mesh, thereby realizing the rotation of the rotating column 404.

[0065] A flipping mechanism 5 is fixedly installed above the weighing device 6. Inside the housing of the flipping mechanism 5, there are a first hydraulic telescopic rod 501 and a second hydraulic telescopic rod 502. The end of the first hydraulic telescopic rod 501 is fixedly connected to the center of the bottom of the first flip plate 503, and the end of the second hydraulic telescopic rod 502 is fixedly connected to the center of the bottom of the second flip plate 504. There is an installation shaft 7 between the first flip plate 503 and the second flip plate 504. The installation shaft 7 is rotatably installed on the upper surface of the housing of the flipping mechanism 5. When the hydraulic telescopic rod is driven to expand and contract, the flip plate is realized to reciprocally flip around the installation shaft 7, and the flipping angle of each flip plate is between 0° and 90°.

[0066] The rotation mechanism 4 and the flipping mechanism 5 of the present invention are both connected to the control computer 1 through the controller. The control computer 1 is used to control the above components to work, thus realizing the automation and intelligence of the information collection of the scrap steel briquette 8 and reducing the labor input. By using the system of the present invention, the image information of six faces of the scrap steel briquette 8 can be effectively collected, and then the collected image information is used to train the convolutional neural network model. More optimally, in the present invention, by optimizing the design of the existing conventional convolutional neural network model, the parameter transfer technology in transfer learning is used to improve the convolutional neural network model, and then the obtained image information is used to train the improved convolutional neural network model, effectively improving the recognition accuracy and speed, and the average recognition time is only 7.5 ms.

[0067] In addition, the system of the present invention can upload the relevant data of the scrap steel briquette 8 to the cloud server for storage, which can be viewed by users at any time. Users can also use the terminal device to take pictures of the scrap steel, upload it to the cloud, and the control computer quickly identifies the relevant identity information of the scrap steel briquette 8. By using the recognition method of the present invention, the rapid, automatic and accurate recognition of the scrap steel identity can be realized, and the entire production process of the scrap steel briquette 8 can be completely supervised and traced, fundamentally eliminating the adulteration behavior.

[0068] The following further describes the present invention in conjunction with specific embodiments.

[0069] Embodiment 1

[0070] Combined with Figures 1-4 As shown, in this embodiment, the scrap steel briquette 8 sample 1 is used. The image information on its surface is collected by the information collection system of the present invention, and the improved calculation model of the present invention is used for automatic processing and recognition. The specific operations are as follows:

[0071] Step A: The monitoring device is turned on, and the scrap steel is packed by the packing machine to obtain the scrap steel briquette 8 sample 1;

[0072] Step B: The scrap steel briquette 8 sample 1 is placed on the flipping mechanism 5. According to the setting of the control computer, the rotation mechanism 4 and the flipping mechanism 5 are automatically driven to cooperate so that the image acquisition device 2 can smoothly collect the images of six faces of the scrap steel briquette 8 sample 1. The images of the six faces collected are as Figure 5 shown.

[0073] At the same time, the weighing device 6 measures the weight of the scrap steel briquette 8 sample 1. The measured weight of the scrap steel briquette 8 sample 1 is 1093 kg, and the obtained image information and weight data are transmitted to the control computer 1;

[0074] Step C: The monitoring device stores the video of the whole process of packing, flipping, weight measurement and image acquisition of the scrap steel briquette 8 taken into the control computer;

[0075] Step D: Remove the current scrap briquette 8 sample 1. The control computer 1 analyzes and processes all the collected relevant data. The control computer 1 uses image processing technology to calculate the size of the scrap briquette 8 sample 1 as 123 cm × 95 cm × 89 cm, and processes the image using the recognition process shown in Figure 4 to predict that the scrap briquette 8 sample 1 is of the steel bar type. After integrating the relevant data of this sample 1, it is completely uploaded to the cloud server 10 to realize the cloud storage of the data information of the briquette sample 1.

[0076] Example 2

[0077] In this example, for the scrap briquette 8 sample 2, the image information on its surface is also collected using the information collection system of the present invention, and the improved calculation model of the present invention is used for automatic processing and recognition. The process is the same as that in Example 1, and the difference is that:

[0078] The measured weight of the scrap briquette 8 sample 2 is 1137 kg;

[0079] The images of the six collected surfaces are as shown in Figure 6 After processing the images, it is obtained that the size of the scrap briquette 8 sample 2 is 102 cm × 91 cm × 88 cm, and the type is galvanized sheet;

[0080] After integrating the relevant data of this sample 2, it is completely uploaded to the cloud server 10 to realize the cloud storage of the data information of the briquette sample 2.

[0081] Example 3

[0082] In this example, for the scrap briquette 8 sample 3, the image information on its surface is also collected using the information collection system of the present invention, and the improved calculation model of the present invention is used for automatic processing and recognition. The process is the same as that in Example 1, and the difference is that:

[0083] The measured weight of the scrap briquette 8 sample 3 is 1137 kg;

[0084] The images of the six collected surfaces are as shown in Figure 7 After processing the images, it is obtained that the size of the scrap briquette 8 sample 3 is 102 cm × 91 cm × 88 cm, and the type is galvanized sheet;

[0085] After integrating the relevant data of this sample 3, it is completely uploaded to the cloud server 10 to realize the cloud storage of the data information of the briquette sample 3.

[0086] In addition, for the image acquisition of the scrap steel briquette 8 samples in Embodiments 1 to 3, a mobile device can also be used to collect pictures and then upload them to the cloud for automated image processing, and all information during the sample production process can also be obtained.

[0087] The present invention has been described in detail above in connection with specific exemplary embodiments. However, it should be understood that various modifications and variations can be made without departing from the scope of the present invention as defined by the appended claims. The detailed description and the drawings should be considered illustrative only and not restrictive. If there are any such modifications and variations, they will all fall within the scope of the present invention described herein. In addition, the background art is intended to illustrate the research and development status and significance of the present technology and is not intended to limit the present invention or the application fields of the present application and the present invention.

[0088] More specifically, although exemplary embodiments of the present invention have been described herein, the present invention is not limited to these embodiments, but includes any and all embodiments that those skilled in the art can recognize from the foregoing detailed description through modifications, omissions, for example, combinations between various embodiments, adaptive changes, and / or substitutions. The limitations in the claims can be broadly interpreted according to the language used in the claims and are not limited to the examples described in the foregoing detailed description or during the implementation of the application. These examples should be considered non-exclusive. Any steps recited in any method or process claim can be executed in any order and are not limited to the order set forth in the claims. Therefore, the scope of the present invention should be determined only by the appended claims and their legal equivalents, rather than by the description and examples given above.

Claims

1. A scrap steel briquette information collection system, characterized in that: It includes an image acquisition device (2), a rotating mechanism (4), a flipping mechanism (5), a weighing device (6) and a monitoring device (9). The scrap steel briquette (8) is placed on the flipping mechanism (5). A weighing device (6) is provided below the flipping mechanism (5). The weighing device (6) together with the flipping mechanism (5) is integrally installed on the rotating mechanism (4). The rotating mechanism (4) drives the scrap steel briquette (8) to rotate. The image acquisition device (2) is arranged on one side of the scrap steel briquette (8) for acquiring the image information of the side of the scrap steel briquette (8). The monitoring device (9) is used for monitoring the process of packing the scrap steel briquette (8) and acquiring the image information. It further includes a control computer (1). The image acquisition device (2), the rotating mechanism (4), the flipping mechanism (5), the weighing device (6) and the monitoring device (9) are all connected to the control computer (1) through a controller. The flipping mechanism (5) is fixedly installed on the weighing device (6). The flipping mechanism (5) includes a mounting shaft (7) and two symmetrically arranged hydraulic telescopic rods. The movable ends of the two hydraulic telescopic rods are respectively fixedly connected to a flap. The flipping angle range of each flap is 0° to 90°. And both flaps are rotatably installed on the mounting shaft (7). The hydraulic telescopic rods are connected to the control computer (1) through a controller.

2. The scrap steel briquetting information acquisition system according to claim 1, characterized in that: It further includes a cloud server (10). The control computer (1) is connected to the cloud server (10), and the cloud server (10) is connected to a mobile device (11).

3. A scrap steel briquette information acquisition system according to any one of claims 1-2, characterized in that: The rotating mechanism (4) includes a driving motor (401), a gear and a rotating column (404). The driving motor (401) is drivingly connected to one end of the rotating column (404) through the gear. The other end of the rotating column (404) is fixedly connected to the weighing device (6). The driving motor (401) and the weighing device (6) are connected to the control computer (1) through a controller.

4. A method for identifying scrap steel briquette information, characterized in that: Use the information acquisition system as described in any one of claims 1 - 3 to acquire the image information of six faces of the scrap steel briquette (8), and then use the acquired image information to train the convolutional neural network model.

5. The method for identifying scrap steel briquette information according to claim 4, characterized in that Specifically, it includes the following steps: Step A: The monitoring device (9) is turned on, and a scrap steel is packed using a packing machine to obtain a scrap steel briquette (8). Step B: Place the scrap steel briquette (8) on the flipping mechanism (5). According to the setting of the control computer (1), automatically drive the rotating mechanism (4) and the flipping mechanism (5) to move successively or simultaneously, so that the image acquisition device (2) can smoothly acquire the images of six faces of the scrap steel briquette (8). At the same time, the weighing device (6) measures the weight of the scrap steel briquette (8), and transmits the acquired image information and weight data to the control computer (1). Step C: The monitoring device (9) stores the video of the whole process of packing, flipping, weight measuring and image acquisition of the scrap steel briquette (8) it shoots into the control computer (1). Step D: The control computer (1) analyzes and processes all the collected relevant data, and then uses the acquired image information to train the convolutional neural network model. Remove the current scrap steel briquette (8) and perform the information acquisition of the next scrap steel briquette (8).

6. A method for identifying scrap steel briquette information according to any one of claims 4-5, characterized in that: The parameter transfer technology in transfer learning is adopted to improve the convolutional neural network model, and then the improved convolutional neural network model is trained with the obtained image information. Subsequently, the trained convolutional neural network model is used at any time to accurately and quickly identify the identity of the steel scrap briquette (8).

Images