Scrap steel grade judging method, device and system

Abstract

The embodiment of the invention provides a waste steel grade judgment method, device and system, and relates to the field of waste steel grade judgment, and the method comprises the steps: obtaining the license plate information of a to-be-detected vehicle, shooting the waste steel of the to-be-detected vehicle, obtaining a waste steel image, obtaining the grade of the waste steel based on the waste steel image, and building the corresponding relation between the license plate information and the grade of the waste steel. The functions of automatic grade inspection and automatic impurity removal of the waste steel are achieved, the subjectivity of manual grade judgment and impurity removal is avoided, and the accuracy and scientificity of judgment are improved.

Description

Technical Field

[0001] The present invention relates to the field of scrap steel grading, and in particular to a scrap steel grading method, device and system. Background Art

[0002] The current scrap steel grading method is usually: the technical center organizes relevant departments to go to the supplier's material yard to lock the quantity and price (on-site grading), and then supervise the loading manually and send it directly to the steelmaking unit. The acceptance personnel of the steelmaking unit check the vehicle number and WeChat photo (the lime mark on the scrap steel freight vehicle) according to the inspection notice issued by the supervisor. Only when there is no obvious difference can it be unloaded. The briquette scrap steel is uniformly sampled. The inspectors check the vehicle number, type, predicted grade and other information according to the inspection entrustment form, and then get on the vehicle (or observe on the ground) to conduct appearance inspection on each vehicle and visually observe the content of the target scattered in the vehicle. Then the dice are rolled to determine the sampling location of the briquette.

[0003] At present, scrap steel inspection relies entirely on manual inspection of scrap steel based on judgment standards, simple measuring tools and the experience of inspectors. Factors such as on-site inspection environment, weather changes, and the inspection personnel's judgment level will cause deviations and influences on the scrap steel inspection results, which can easily lead to disputes. Due to limited conditions, the inspection process cannot trace the quality of each piece of scrap steel, resulting in no evidence for quality objections, risk loopholes, and increased difficulty in prevention and control. Summary of the invention

[0004] The purpose of the present invention is to provide a scrap steel grading method, device and system, which can realize automatic grading of scrap steel, avoid the subjectivity of manual grading and improve the scientific nature of the determination.

[0005] In order to achieve the above purpose, the technical solution adopted by the embodiment of the present invention is as follows:

[0006] In a first aspect, an embodiment of the present invention provides a scrap steel grading method, which is applied to a processor of a scrap steel grading system, and the method includes:

[0007] Obtain the license plate information of the vehicle to be detected;

[0008] Photographing the scrap steel of the vehicle to be inspected to obtain a scrap steel image;

[0009] The scrap steel grade is obtained based on the scrap steel image, and a corresponding relationship between the license plate information and the scrap steel grade is established.

[0010] In an optional embodiment, the scrap steel grading system includes a magnetic crane, a grabber or a packer, and the scrap steel grading system also includes a gun camera. The magnetic crane is used to adsorb bulk materials, the grabber is used to grab scrap steel briquettes, and the packer is used to unpack the scrap steel briquettes. The step of photographing the scrap steel of the vehicle to be detected to obtain a scrap steel image includes:

[0011] When it is detected that the magnetic crane leaves the compartment area of ​​the vehicle to be detected, the gun camera is controlled to shoot the scrap steel to obtain an image of the scrap steel;

[0012] When it is detected that the grabber leaves the compartment area of ​​the vehicle to be detected, the gun camera is controlled to shoot the scrap steel to obtain a scrap steel image;

[0013] After detecting that the unpacking machine has finished unpacking the scrap steel briquettes, the gun camera is controlled to photograph the scrap steel to obtain a scrap steel image.

[0014] In an optional embodiment, the method further comprises:

[0015] Using a filter to remove noise in the scrap steel image to obtain a first scrap steel image;

[0016] Processing the first scrap steel image by an edge detection algorithm to obtain a second scrap steel image;

[0017] performing smoothing processing on the second scrap steel image to obtain a third scrap steel image;

[0018] Processing the third scrap steel image through an opening operation and a closing operation to obtain a fourth scrap steel image;

[0019] Processing the fourth scrap steel image based on histogram equalization to obtain a target scrap steel image;

[0020] The step of obtaining the scrap steel grade based on the scrap steel image and establishing a corresponding relationship between the license plate information and the scrap steel grade comprises:

[0021] The scrap steel grade is obtained based on the target scrap steel image, and a corresponding relationship between the license plate information and the scrap steel grade is established.

[0022] In an optional embodiment, the system further includes a fill light, and the fill light is used to provide fill light when photographing the scrap steel image. The method further includes:

[0023] Get the current ambient light intensity;

[0024] Comparing the current ambient light intensity with a preset ambient light intensity;

[0025] Based on the comparison result, the fill light is controlled.

[0026] In an optional embodiment, the step of obtaining the scrap steel grade based on the target scrap steel image comprises:

[0027] Segmenting the target scrap steel image to obtain a plurality of scrap steel sub-images;

[0028] For each scrap steel sub-image, obtaining a category of the scrap steel from the scrap steel sub-image;

[0029] Based on the categories of all the scrap steel sub-images, determining the proportion of each scrap steel category in the scrap steel image;

[0030] Based on the proportion of each type of scrap steel, the scrap steel grade of the target scrap steel image is determined.

[0031] In an optional embodiment, the step of acquiring the category of the scrap steel from each scrap steel sub-image comprises:

[0032] Inputting the scrap steel sub-image into a trained scrap steel recognition model to obtain high-level features of the scrap steel sub-image, wherein the high-level features include a bounding box, a segmentation mask, and a category probability of a scrap steel monomer target;

[0033] Determine the underlying features of the scrap steel sub-image, wherein the underlying features include edge thickness of the scrap steel monomer, scrap steel texture, scrap steel color, and scrap steel shape;

[0034] Fusing the bottom-level features and the high-level features to obtain fused features;

[0035] Based on the fusion features, the scrap steel type of the scrap steel sub-image is obtained.

[0036] In an optional implementation manner, the step of fusing the bottom-level features with the high-level features to obtain fused features includes:

[0037] Determine the bounding box, segmentation mask, category probability, edge thickness of the scrap steel monomer, scrap steel texture, scrap steel color, and weight of the scrap steel shape of the scrap steel monomer target respectively;

[0038] Calculating a first product of a bounding box of the scrap steel monomer target and a corresponding weight;

[0039] Calculating a second product of the segmentation mask and the corresponding weight;

[0040] Calculating a third product of the class probability and the corresponding weight;

[0041] Calculating a fourth product of the edge thickness of the scrap steel monomer and the corresponding weight;

[0042] Calculating a fifth product of the scrap steel texture and the corresponding weight of the scrap steel monomer;

[0043] Calculating a sixth product of the scrap steel color and the corresponding weight of the scrap steel monomer;

[0044] Calculating the seventh product of the scrap steel shape and the corresponding weight of the scrap steel monomer;

[0045] The bottom-level features and the high-level features are fused based on the first product, the second product, the third product, the fourth product, the fifth product, the sixth product, and the seventh product to obtain fused features.

[0046] In an optional implementation manner, the step of fusing the bottom-level features with the high-level features to obtain fused features includes:

[0047] Inputting the bottom-level features and the high-level features into the feature pyramid network for feature splicing to determine a splicing feature map;

[0048] Inputting the concatenated feature map into the convolution layer of the feature pyramid network, performing convolution processing on the concatenated feature map, and determining a first target feature map;

[0049] Inputting the first target feature map into the pooling layer of the feature pyramid network, performing global average pooling processing on the first target feature map, and determining a second target feature map;

[0050] Inputting the second target feature map into the activation layer of the feature pyramid network, performing activation processing on the second target feature map, and determining a first target weight and a second target weight;

[0051] The first target weight and the second target weight are input into the fusion layer of the feature pyramid network, the bottom-level features are weighted by the first target weight and the high-level features are weighted by the second target weight to determine the fusion feature.

[0052] In a second aspect, an embodiment of the present invention provides a scrap steel grading device, the device comprising:

[0053] An acquisition module is used to obtain the license plate information of the vehicle to be detected;

[0054] A shooting module, used for shooting the scrap steel of the vehicle to be detected to obtain a scrap steel image;

[0055] The grade determination module is used to obtain the scrap steel grade based on the scrap steel image and establish a corresponding relationship between the license plate information and the scrap steel grade.

[0056] In a third aspect, an embodiment of the present invention provides a scrap steel grading system, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the scrap steel grading method when executing the computer program.

[0057] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the scrap steel grading method when executed by a processor.

[0058] The present invention has the following beneficial effects:

[0059] The present invention obtains the license plate information of the vehicle to be detected, photographs the scrap steel of the vehicle to be detected, obtains the scrap steel image, obtains the scrap steel grade based on the scrap steel image, and establishes the corresponding relationship between the license plate information and the scrap steel grade. The automatic grade inspection and automatic impurity removal functions of scrap steel are realized, the subjectivity of manual grade determination and impurity removal is avoided, and the accuracy and scientificity of the determination are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0061] Figure 1 A block diagram of a scrap steel grading system provided by an embodiment of the present invention;

[0062] Figure 2 One of the flow diagrams of a scrap steel grading method provided in an embodiment of the present invention;

[0063] Figure 3 A second flow chart of a scrap steel grading method provided by an embodiment of the present invention;

[0064] Figure 4 A third flow chart of a scrap steel grading method provided by an embodiment of the present invention;

[0065] Figure 5 A fourth flow chart of a scrap steel grading method provided in an embodiment of the present invention;

[0066] Figure 6 A fifth flow chart of a scrap steel grading method provided in an embodiment of the present invention;

[0067] Figure 7 A schematic structural diagram of a scrap steel grading device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to make the purpose, 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 in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0069] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0070] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0071] In the description of the present invention, it should be noted that if the terms "upper", "lower", "inside", "outside", etc. appear to indicate an orientation or position relationship, they are based on the orientation or position relationship shown in the accompanying drawings, or are the orientation or position relationship in which the product of the invention is usually placed when used. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0072] In addition, the terms “first”, “second”, etc., if used, are merely used to distinguish between the descriptions and should not be understood as indicating or implying relative importance.

[0073] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0074] After extensive research by the inventors, it was found that the current scrap steel grading method is usually: the technical center organizes relevant departments to go to the supplier's material yard to lock the quantity and price (on-site grading), and then supervises the loading manually and sends it directly to the steelmaking unit. The acceptance personnel of the steelmaking unit check the vehicle number and WeChat photo (the lime mark on the scrap steel freight vehicle) according to the inspection notice issued by the supervisor. Only when there is no obvious difference can the vehicle be unloaded. The briquette scrap steel is uniformly sampled. The inspectors check the vehicle number, variety, and predicted grade according to the inspection entrustment form, and then get on the vehicle (or observe on the ground) to conduct an appearance inspection of each vehicle, and visually observe the content of the target scattered in the vehicle. The briquette sampling location is determined by rolling the dice. Due to the constraints and limitations of human factors, the representativeness and randomness of the sampling are not high.

[0075] At present, random sampling of briquetted scrap steel is carried out at the 3# scrap steel warehouse (2 briquettes of steel bars are randomly sampled for unpacking in each vehicle due to their good permeability, and 4 briquettes are randomly sampled in each vehicle for other types). Due to the limited location of the 3# scrap steel warehouse, there is only one entrance and exit channel, which causes traffic congestion, and the loading and unloading time of scrap steel sampling is long, which not only has low detection efficiency, but also poses safety hazards.

[0076] At present, scrap steel grading relies entirely on manual inspection of scrap steel based on judgment standards, simple measuring tools and the experience of inspectors. Factors such as on-site inspection environment, weather changes, and the inspection personnel's judgment level will cause deviations and influences on the scrap steel inspection results, which can easily lead to disputes. Due to limited conditions, the inspection process cannot trace the quality of each piece of scrap steel, resulting in no evidence for quality objections, risk loopholes, and increased difficulty in prevention and control.

[0077] Under the premise that the output of molten iron in blast furnaces is limited, the Thick Plate Special Steel Division and the Silicon Steel Thin Plate Division have increased steel production by reducing the iron-steel ratio and consuming more scrap steel. In 2023, the average number of vehicles entering the factory with briquetted scrap steel will be 47.5 vehicles / day, and the average number of vehicles entering the factory with bulk scrap steel will be 9.7 vehicles / day. If the scrap steel vehicles are loaded according to the standard load, the number of vehicles entering the factory will reach 139 vehicles / day. The company's scrap steel inspection workload has increased significantly, and manpower is severely limited. Suppliers use high-side vehicles to transport scrap steel, which is not conducive to the steel grabber to randomly inspect the scrap steel at the bottom of the vehicle. The inspectors need to inspect the quality of the upper layer, notify the user to unload the scrap steel briquettes on the upper layer, and then return to the scrap steel warehouse to inspect the quality of the scrap steel on the lower layer. The scrap steel inspection cycle is long, and the scrap steel briquettes are placed irregularly, which is not conducive to random inspections, and there is also a risk of missed inspections.

[0078] In view of the discovery of the above problems, this embodiment provides a scrap steel grading method, device and system, which can obtain the license plate information of the vehicle to be detected, take pictures of the scrap steel of the vehicle to be detected, obtain the scrap steel image, obtain the scrap steel grade based on the scrap steel image, and establish the corresponding relationship between the license plate information and the scrap steel grade. The automatic grading and automatic deduction of impurities of scrap steel are realized, avoiding the subjectivity of manual grading and deduction of impurities, and improving the accuracy and scientificity of the judgment. The scheme provided by this embodiment is described in detail below.

[0079] Please refer to Figure 1 , Figure 1 1 is a schematic diagram of the structure of the scrap steel grading system 100 provided in an embodiment of the present invention. The scrap steel grading system 100 may also include Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown. Figure 1 Each component shown in the figure can be implemented by hardware, software or a combination thereof.

[0080] The scrap steel grading system 100 includes a scrap steel grading device 110 , a memory 120 and a processor 130 .

[0081] The memory 120 and the processor 130 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines. The scrap steel grading device 110 includes at least one software function module that can be stored in the memory 120 in the form of software or firmware or fixed in the operating system (OS) of the scrap steel grading system 100. The processor 130 is used to execute the executable modules stored in the memory 120, such as the software function modules and computer programs included in the scrap steel grading device 110.

[0082] The memory 120 may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc. The memory 120 is used to store a program, and the processor 130 executes the program after receiving an execution instruction.

[0083] Please refer to Figure 2 , Figure 2 For application Figure 1 A flowchart of a scrap steel grading method of the scrap steel grading system 100 is provided, and the method including various steps is described in detail below.

[0084] S201: Obtaining the license plate information of the vehicle to be detected.

[0085] S202: Photographing the scrap steel of the vehicle to be inspected to obtain a scrap steel image.

[0086] S203: Obtaining the scrap steel grade based on the scrap steel image, and establishing a corresponding relationship between the license plate information and the scrap steel grade.

[0087] The scrap steel grading system may include a processor for processing and analyzing the received scrap steel image to obtain the scrap steel grade corresponding to the scrap steel image.

[0088] The scrap steel grading system may further include an entry recognition device for recognizing the license plate information of the vehicle to be detected and obtaining the license plate information of the vehicle to be detected.

[0089] The scrap steel grading system may also include a camera or monitoring device that combines a ball camera and a gun camera to detect the driving path of the vehicle to be inspected and monitor the model of the vehicle to be inspected. After the vehicle to be inspected enters the site, the model of the vehicle to be inspected is determined based on the camera or monitoring device that combines a ball camera and a gun camera. The processor of the scrap steel grading system determines the specified inspection location of the vehicle to be inspected based on the model of the vehicle to be inspected.

[0090] The scrap steel grading system may further include a display, which may be used to display a prescribed inspection location of the vehicle to be inspected, so that the driver can know the prescribed inspection location and drive to the prescribed inspection location.

[0091] The scrap steel system has the functions of unpacking and briquetting, bulk scrap steel grade identification, and automatic identification of sundries and garbage, and uploads the results to the processor in a timely manner. The scrap steel system has the function of queuing and calling numbers for scrap steel vehicles. The system sorts the vehicles entering the scrap steel plant and allocates quality inspection lanes according to the relevant data of the MES system and the existing queuing and calling sorting rules. The scrap steel system has the function of auto-focusing and taking pictures, automatically tracking the location of the magnetic crane / grab machine, and providing high coverage, low overlap, and high-quality images for the algorithm model. The scrap steel system has the function of identifying and alarming dangerous objects such as oily parts, closed containers, and explosives, and prompts on the display.

[0092] The camera or monitoring equipment combining the ball camera and the gun camera can also identify the overall image of the vehicle to be inspected, and send the overall image of the vehicle to be inspected to the processor of the scrap steel grading system, so that the processor of the scrap steel grading system can identify the dangerous goods, closed containers and seriously adulterated false information in the overall image according to the overall image of the vehicle to be inspected, and transmit the corresponding recognition results and the license plate information of the vehicle to be inspected to the display for display.

[0093] The scrap steel grading system also includes a voice broadcasting device, which can broadcast information such as the specified inspection location of the vehicle to be inspected, dangerous goods in the overall image, closed containers, and seriously adulterated and false information.

[0094] The scrap steel grading system may further include a suction cup. When the scrap steel on the vehicle to be inspected is bulk material, after the vehicle to be inspected enters the specified inspection location, the camera or monitoring equipment combined with the ball camera and the gun camera can obtain the scrap steel image on the vehicle to be inspected layer by layer, that is, first obtain the scrap steel image of the top layer, then suck away the scrap steel on the top layer through the suction cup, and then photograph the scrap steel on the lower layer to obtain the image, and so on, until the scrap steel images of each layer of the scrap steel on the vehicle to be inspected are collected.

[0095] When the scrap steel of the vehicle to be inspected includes bulk material and scrap steel briquettes, several scrap steel briquettes can be extracted for photographing, the bulk scrap steel can be adsorbed layer by layer by a suction cup and photographed, and the scrap steel grade can be determined based on the scrap steel images of the scrap steel briquettes and the bulk scrap steel images.

[0096] When the scrap steel on the vehicle to be inspected is scrap steel briquettes, the number of layers of the scrap steel briquettes is randomly inspected and the results are displayed on the on-site display. The driver of the bag grabber grabs the briquettes according to the position indicated on the large screen, unpacks them, takes photos, and automatically completes the scrap steel grading.

[0097] After collecting the scrap steel image, each piece of scrap steel in the image needs to be segmented for subsequent identification. The trained recognition network can be used to identify the scrap steel in the scrap steel image.

[0098] In one example, the scrap steel image may be an image captured by a camera in real time, or one or more frames of images may be obtained from a video captured by a camera in real time.

[0099] In another example, the scrap steel image may be an image input by a user or selected by a user. For example, the user inputs an image to be identified on a display interface of the terminal, or selects a scrap steel image. Alternatively, the server receives the scrap steel image input by a user or selected by a user sent by the terminal.

[0100] The scrap steel category of each pixel in the scrap steel area in the scrap steel image is counted, so that the scrap steel grade of the whole vehicle can be evaluated, and the corresponding relationship between the license plate information of the vehicle to be detected and the scrap steel grade can be established, so that the scrap steel can be traced.

[0101] There are many ways to photograph the scrap steel of the vehicle to be inspected and obtain the scrap steel image. In one implementation, Figure 3 As shown, the following steps are included:

[0102] The scrap steel grading system includes a magnetic crane, a grabber or a packer. The scrap steel grading system also includes a gun camera, the magnetic crane is used to absorb bulk materials, the grabber is used to grab scrap steel briquettes, and the packer is used to unpack the scrap steel briquettes.

[0103] S301: When it is detected that the magnetic crane leaves the compartment area of ​​the vehicle to be detected, the gun camera is controlled to shoot the scrap steel to obtain a scrap steel image.

[0104] S302: When it is detected that the grabber leaves the compartment area of ​​the vehicle to be inspected, the gun camera is controlled to shoot the scrap steel to obtain a scrap steel image.

[0105] S303: After detecting that the unpacking machine has finished unpacking the scrap steel briquettes, controlling the gun camera to photograph the scrap steel to obtain a scrap steel image.

[0106] In order to avoid interference factors in the collected scrap steel images, when it is detected that the magnetic crane leaves the car area of ​​the vehicle to be inspected, the machine camera is controlled to shoot the scrap steel to obtain the scrap steel image. When it is detected that the grabber leaves the car area of ​​the vehicle to be inspected, the gun camera is controlled to shoot the scrap steel to obtain the scrap steel image. When it is detected that the unpacking machine has finished unpacking the scrap steel briquettes, the gun camera is controlled to shoot the scrap steel to obtain the scrap steel image.

[0107] When shooting scrap steel images, the changes in the scrap steel area in the carriage can be monitored at all times based on the carriage scrap steel area detection model, and the suction cup tracking model can track the suction cup. When the suction cup is tracked to enter the scrap steel area in the carriage and perform the unloading operation, the scrap steel area feature comparison model is used to calculate the size of the scrap steel area sucked by the suction cup. According to the size of the sucked scrap steel area and combined with the focusing algorithm, the focal length adjustment ratio of the camera is determined, and the focal length adjustment ratio is passed to the camera control model, that is, the camera ball camera or gun camera is controlled to collect photos of the sucked scrap steel area to obtain a scrap steel image, thereby obtaining a clear scrap steel image, and the captured scrap steel image is wider, and the scrap steel image is closer to the actual scrap steel grade distribution.

[0108] In order to improve the efficiency of scrap steel image processing, the captured scrap steel images can be preprocessed first, such as Figure 4As shown, the following steps are included:

[0109] S401: Use a filter to remove noise in the scrap steel image to obtain a first scrap steel image.

[0110] S402: Processing the first scrap steel image by using an edge detection algorithm to obtain a second scrap steel image.

[0111] S403: Smoothing the second scrap steel image to obtain a third scrap steel image.

[0112] S404: Processing the third scrap steel image through an opening operation and a closing operation to obtain a fourth scrap steel image.

[0113] S405: Processing the fourth scrap steel image based on histogram equalization to obtain a target scrap steel image.

[0114] Use filters to remove noise from images; extract important features through edge detection algorithms; reduce details in images and highlight main features through smoothing; remove small noise and connected objects through opening and closing operations; use histogram equalization to enhance the contrast of images.

[0115] The way to remove noise from scrap steel images can be based on the mean filtering method, replacing the value of the central pixel with the average value of the neighboring pixels. You can also use median filtering, replacing the value of the central pixel with the median of the neighboring pixels. Use Gaussian filtering to remove noise from scrap steel images, using a Gaussian kernel for convolution and weighted averaging of neighboring pixels. Use bilateral filtering to remove noise from scrap steel images, combining the weighted average of spatial distance and pixel intensity differences. Use non-local mean denoising to use information from similar areas in the image for denoising. Use wavelet transform denoising to decompose the image into sub-bands of different frequencies using wavelet transform, and then threshold each sub-band. Use deep learning methods for denoising, using deep neural networks such as CNN, U-Net, etc. to learn the mapping from noisy images to clean images.

[0116] The edge detection algorithm can perform edge detection on the first scrap steel image based on the Sobel operator, perform edge detection on the first scrap steel image based on the Prewitt operator, perform edge detection on the first scrap steel image based on the Canny edge detection, perform edge detection on the first scrap steel image based on the Laplacian operator, and perform edge detection on the first scrap steel image based on the Roberts operator.

[0117] Image smoothing is a common image preprocessing technology used to reduce noise and details in images and make the image look smoother. Mean filtering, Gaussian filtering, median filtering, bilateral filtering, and non-local mean denoising can be used to smooth the second scrap steel image.

[0118] In addition to the above-mentioned methods of using a filter to remove noise in the scrap steel image to obtain a first scrap steel image, processing the first scrap steel image through an edge detection algorithm to obtain a second scrap steel image, smoothing the second scrap steel image to obtain a third scrap steel image, processing the third scrap steel image through opening and closing operations to obtain a fourth scrap steel image, and processing the fourth scrap steel image based on histogram equalization to obtain a target scrap steel image, a filter can also be used to remove noise in the scrap steel image, process the scrap steel image through an edge detection algorithm, smooth the scrap steel image, process the scrap steel image through opening and closing operations, and process the scrap steel image based on histogram equalization to finally obtain a target scrap steel image.

[0119] The scrap steel image can also be preprocessed in the following way: the acquired scrap steel image includes scrap steel, scrap steel transport vehicles and the surrounding environment. In order to improve efficiency and reduce system overhead, the scrap steel transport vehicles and the surrounding environment in the image can be separated before the next operation, and only the scrap steel is subsequently processed.

[0120] The purpose of preprocessing scrap steel images is to eliminate irrelevant information in the image, restore useful real information, enhance the detectability of relevant information, and simplify data to the maximum extent, thereby improving the reliability of feature extraction, image segmentation, matching, and recognition. In this project, due to the dark color of scrap steel itself, the contrast of the captured image is low. In order to ensure the reliability of subsequent image recognition, the project uses histogram equalization for image enhancement and enhances the before-after contrast.

[0121] The scrap steel grade is obtained based on the target scrap steel image, and the corresponding relationship between the license plate information and the scrap steel grade is established.

[0122] In order to ensure that the scrap steel images are clear and easy to identify, Figure 5 As shown, the following steps are included:

[0123] S501: Obtain the current ambient light intensity.

[0124] S502: Compare the current ambient light intensity with the preset ambient light intensity.

[0125] S503: Based on the comparison result, the fill light is controlled.

[0126] When the current ambient light intensity is greater than the preset ambient light intensity, a first difference between the current ambient light intensity and the preset ambient light intensity is calculated, and the fill light is compensated based on the first difference, that is, the intensity of the fill light is increased.

[0127] When the current ambient light intensity is less than the preset ambient light intensity, calculate the second difference between the current ambient light intensity and the preset ambient light intensity, and compensate the fill light based on the second difference, that is, reduce the intensity of the light of the fill light.

[0128] There are multiple ways to obtain the scrap steel grade based on the target scrap steel image. In one implementation, as Figure 6 shown, it includes the following steps:

[0129] S601: Segment the target scrap steel image to obtain multiple scrap steel sub-images.

[0130] S602: For each scrap steel sub-image, obtain the category of the scrap steel from the scrap steel sub-image.

[0131] S603: Based on the categories of all scrap steel sub-images, determine the proportion of each scrap steel category in the scrap steel image.

[0132] S604: Based on the proportion of each scrap steel category, determine the scrap steel grade of the target scrap steel image.

[0133] After the scrap steel image can be collected, each piece of scrap steel in the image needs to be segmented for subsequent identification. In this embodiment, an instance segmentation artificial neural network model is used to identify and frame the target to be recognized in the image. Calculate the size of each target to be recognized, and classify and grade the scrap steel according to the size of the target to be recognized. Specifically, it also includes: setting size indicators for different types of scrap steel; for example: the thickness of scrap steel A is B, the shear material is C, the thickness of the light and thin pressed block is <D, the diameter of the steel bar is E...

[0134] Set different labels for different types of scrap steel; input the scrap steel type labels, and different scrap steels use different labels. Use the scrap steel grading artificial neural network model to identify the types of the segmented targets to be recognized according to the size indicators of the scrap steel. Calculate the proportion of each scrap steel according to the proportion of the label represented by each scrap steel in all the labels in the image, and conduct an overall evaluation of all the scrap steel according to the proportion of each scrap steel.

[0135] The method of segmenting the target scrap steel image can be to identify the target scrap steel image based on the target detection model to obtain each detection frame, and segment the target scrap steel image based on each detection frame to obtain each scrap steel sub-image. Or use per-pixel image segmentation means to segment the target scrap steel image to obtain multiple scrap steel sub-images.

[0136] Since scrap steel comes in many types and shapes, it is difficult to determine the category of scrap steel only by appearance and color. Therefore, in order to improve the accuracy of scrap steel recognition, the scrap steel sub-image is input into the trained scrap steel recognition model to obtain the high-level features of the scrap steel sub-image, where the high-level features include the bounding box, segmentation mask and category probability of the scrap steel monomer target; determine the underlying features of the scrap steel sub-image, where the underlying features include the edge thickness of the scrap steel monomer, the scrap steel texture, the scrap steel color and the scrap steel shape; fuse the underlying features and the high-level features to obtain the fused features; based on the fused features, obtain the scrap steel type of the scrap steel sub-image.

[0137] One implementation method of fusing the bottom-level features and the high-level features to obtain the fused features may be:

[0138] Determine the bounding box, segmentation mask, category probability, edge thickness of the scrap steel monomer, scrap steel texture, scrap steel color and weight of the scrap steel shape of the scrap steel monomer target respectively; calculate the first product of the bounding box of the scrap steel monomer target and the corresponding weight; calculate the second product of the segmentation mask and the corresponding weight; calculate the third product of the category probability and the corresponding weight; calculate the fourth product of the edge thickness of the scrap steel monomer and the corresponding weight; calculate the fifth product of the scrap steel texture of the scrap steel monomer and the corresponding weight; calculate the sixth product of the scrap steel color of the scrap steel monomer and the corresponding weight; calculate the seventh product of the scrap steel shape of the scrap steel monomer and the corresponding weight; fuse the underlying features and high-level features based on the first product, the second product, the third product, the fourth product, the fifth product, the sixth product and the seventh product to obtain the fused features.

[0139] The scrap steel identification model is based on deep learning and big data technology and needs to be combined with human experience to convert human experience into a language that the algorithm model can understand, that is, to add human expert knowledge to the image data. Therefore, the calibration of scrap steel grade and scrap steel impurity content requires a large amount of data library support.

[0140] The fused features can be used to identify scrap steel based on Fast R-CNN to obtain the category to which the scrap steel belongs.

[0141] Another implementation method of fusing the bottom-level features and the high-level features to obtain the fused features may be:

[0142] The bottom-level features and the high-level features are input into the feature pyramid network for feature splicing to determine a splicing feature map; the splicing feature map is input into the convolution layer of the feature pyramid network, the splicing feature map is convolved to determine a first target feature map; the first target feature map is input into the pooling layer of the feature pyramid network, the first target feature map is globally averaged pooled to determine a second target feature map; the second target feature map is input into the activation layer of the feature pyramid network, the second target feature map is activated to determine a first target weight and a second target weight; the first target weight and the second target weight are input into the fusion layer of the feature pyramid network, the bottom-level features are weighted using the first target weight and the high-level features are weighted using the second target weight to determine the fusion feature.

[0143] Please refer to Figure 7 The embodiment of the present invention also provides a method for applying Figure 1 The scrap steel grading device 110 of the scrap steel grading system 100 includes:

[0144] The acquisition module 111 is used to acquire the license plate information of the vehicle to be detected;

[0145] A shooting module 112 is used to shoot the scrap steel of the vehicle to be detected to obtain a scrap steel image;

[0146] The grade determination module 113 is used to obtain the scrap steel grade based on the scrap steel image and establish a corresponding relationship between the license plate information and the scrap steel grade.

[0147] The present invention also provides a scrap steel grading system 100, which includes a processor 130 and a memory 120. The memory 120 stores computer executable instructions, and when the computer executable instructions are executed by the processor 130, the scrap steel grading method is implemented.

[0148] The embodiment of the present invention further provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by the processor 130, the scrap steel grading method is implemented.

[0149] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, program segment or a part of the code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0150] In addition, each functional module in each embodiment of the present invention can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

[0151] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0152] The above are only various embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for grading scrap steel, characterized in that: A processor applied to a scrap steel grading system, the method comprising: Obtain the license plate information of the vehicle to be detected; Photographing the scrap steel of the vehicle to be inspected to obtain a scrap steel image; The scrap steel grade is obtained based on the scrap steel image, and a corresponding relationship between the license plate information and the scrap steel grade is established.

2. The method according to claim 1, characterized in that The scrap steel grading system includes a magnetic crane, a grabber or a packer, and the scrap steel grading system also includes a gun camera. The magnetic crane is used to adsorb bulk materials, the grabber is used to grab scrap steel briquettes, and the packer is used to unpack the scrap steel briquettes. The step of photographing the scrap steel of the vehicle to be detected to obtain a scrap steel image includes: When it is detected that the magnetic crane leaves the compartment area of ​​the vehicle to be detected, the gun camera is controlled to shoot the scrap steel to obtain an image of the scrap steel; When it is detected that the grabber leaves the compartment area of ​​the vehicle to be detected, the gun camera is controlled to shoot the scrap steel to obtain a scrap steel image; After detecting that the unpacking machine has finished unpacking the scrap steel briquettes, the gun camera is controlled to photograph the scrap steel to obtain a scrap steel image.

3. The method according to claim 1 or 2, characterized in that: The method further comprises: Using a filter to remove noise in the scrap steel image to obtain a first scrap steel image; Processing the first scrap steel image by an edge detection algorithm to obtain a second scrap steel image; performing smoothing processing on the second scrap steel image to obtain a third scrap steel image; Processing the third scrap steel image through an opening operation and a closing operation to obtain a fourth scrap steel image; Processing the fourth scrap steel image based on histogram equalization to obtain a target scrap steel image; The step of obtaining the scrap steel grade based on the scrap steel image and establishing a corresponding relationship between the license plate information and the scrap steel grade comprises: The scrap steel grade is obtained based on the target scrap steel image, and a corresponding relationship between the license plate information and the scrap steel grade is established.

4. The method according to claim 1, characterized in that: The system further includes a fill light, and the fill light is used to fill light when shooting the scrap steel image. The method further includes: Get the current ambient light intensity; Comparing the current ambient light intensity with a preset ambient light intensity; Based on the comparison result, the fill light is controlled.

5. The method according to claim 3, characterized in that: The step of obtaining the scrap steel grade based on the target scrap steel image comprises: Segmenting the target scrap steel image to obtain a plurality of scrap steel sub-images; For each scrap steel sub-image, obtaining a category of the scrap steel from the scrap steel sub-image; Based on the categories of all the scrap steel sub-images, determining the proportion of each scrap steel category in the scrap steel image; Based on the proportion of each type of scrap steel, the scrap steel grade of the target scrap steel image is determined.

6. The method according to claim 5, characterized in that The step of obtaining the category of the scrap steel from each scrap steel sub-image comprises: Inputting the scrap steel sub-image into a trained scrap steel recognition model to obtain high-level features of the scrap steel sub-image, wherein the high-level features include a bounding box, a segmentation mask, and a category probability of a scrap steel monomer target; Determine the underlying features of the scrap steel sub-image, wherein the underlying features include edge thickness of the scrap steel monomer, scrap steel texture, scrap steel color, and scrap steel shape; Fusing the bottom-level features and the high-level features to obtain fused features; Based on the fusion features, the scrap steel type of the scrap steel sub-image is obtained.

7. The method according to claim 6, characterized in that The step of fusing the bottom-level features and the high-level features to obtain fused features includes: Determine the bounding box, segmentation mask, category probability, edge thickness of the scrap steel monomer, scrap steel texture, scrap steel color, and weight of the scrap steel shape of the scrap steel monomer target respectively; Calculating a first product of a bounding box of the scrap steel monomer target and a corresponding weight; Calculating a second product of the segmentation mask and the corresponding weight; Calculating a third product of the class probability and the corresponding weight; Calculating a fourth product of the edge thickness of the scrap steel monomer and the corresponding weight; Calculating a fifth product of the scrap steel texture and the corresponding weight of the scrap steel monomer; Calculating a sixth product of the scrap steel color and the corresponding weight of the scrap steel monomer; Calculating the seventh product of the scrap steel shape and the corresponding weight of the scrap steel monomer; The bottom-level features and the high-level features are fused based on the first product, the second product, the third product, the fourth product, the fifth product, the sixth product, and the seventh product to obtain fused features.

8. The method according to claim 6, characterized in that The step of fusing the bottom-level features with the high-level features to obtain fused features includes: Inputting the bottom-level features and the high-level features into a feature pyramid network for feature splicing to determine a splicing feature map; Inputting the concatenated feature map into the convolution layer of the feature pyramid network, performing convolution processing on the concatenated feature map, and determining a first target feature map; Inputting the first target feature map into the pooling layer of the feature pyramid network, performing global average pooling processing on the first target feature map, and determining a second target feature map; Inputting the second target feature map into the activation layer of the feature pyramid network, performing activation processing on the second target feature map, and determining a first target weight and a second target weight; The first target weight and the second target weight are input into the fusion layer of the feature pyramid network, the bottom-level features are weighted by the first target weight and the high-level features are weighted by the second target weight to determine the fusion feature.

9. A scrap steel grading device, characterized in that: The device comprises: An acquisition module is used to obtain the license plate information of the vehicle to be detected; A shooting module, used for shooting the scrap steel of the vehicle to be detected to obtain a scrap steel image; The grade determination module is used to obtain the scrap steel grade based on the scrap steel image and establish a corresponding relationship between the license plate information and the scrap steel grade.

10. A scrap steel grading system, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method according to any one of claims 1 to 8 when executing the computer program.

Images