Quantitative detection method and device for mixed coal components based on machine vision and medium

Abstract

The invention relates to a quantitative detection method and device for mixed coal components based on machine vision and a medium, and belongs to the technical field of mixed coal component recognizing.The method comprises the steps that a mixed coal image of to-be-detected mixed coal is obtained, and the to-be-detected mixed coal comprises multiple kinds of single coal; inputting the mixed coal image into a pre-trained mixed coal quantitative identification model to obtain a quantitative detection result of the to-be-detected mixed coal, the quantitative detection result including the coal type and content of each single coal in the to-be-detected mixed coal; wherein the mixed coal quantitative recognition model is obtained by training a plurality of target mixed coal sample images, and each target mixed coal sample image corresponds to the coal type and content of each single coal in a plurality of single coals. The method has the effect of improving the quantitative detection efficiency of the components of the mixed coal.

Description

Technical Field

[0001] The present invention relates to the technical field of blended coal component identification, and in particular, to a method, device and medium for quantitatively detecting blended coal components based on machine vision. Background Art

[0002] In the field of thermal power generation, coal is the main source of energy supply, and its cost-effectiveness and combustion efficiency are directly related to the operating cost of the power plant and the sustainability of power production. To optimize the coal combustion cost, power plants usually adopt a strategy of blending the designed coal type with a non-designed coal type with a lower price for combustion. This blending method aims to balance the coal combustion cost and the stable operation of coal-fired equipment, ensuring that while meeting the power demand, the operating cost is reduced as much as possible.

[0003] However, the proportion of different components in the blended coal quality has a significant impact on the operating stability of coal-fired equipment. Especially when the proportion of the non-designed coal type exceeds a certain critical value, it may cause failures of coal-fired equipment, such as increased wear, ash fouling and coking, etc., which will in turn affect the power generation efficiency and equipment life. Therefore, accurately controlling the proportion of each component in the blended coal quality to ensure it is below the critical value is the key to ensuring the stable operation of coal-fired equipment and reducing the economic cost of coal combustion.

[0004] Traditional methods for detecting the content of blended coal quality mainly rely on regular sampling organized by power plants and laboratory chemical tests to determine. Although this method can provide relatively accurate coal quality component data, it has obvious limitations. On the one hand, the processes of regular sampling and laboratory tests are time-consuming and cannot reflect the changes in coal quality components in real time, making it difficult to adjust the blending ratio of each coal type in a timely manner. On the other hand, frequent sampling and testing will also increase the operating cost and manpower burden of the power plant. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method, device and medium for quantitatively detecting blended coal components based on machine vision, aiming to solve at least one of the above technical problems.

[0006] The technical solution of the present invention to solve the above technical problems is as follows:

[0007] In the first aspect, the present application provides a method for quantitatively detecting blended coal components based on machine vision, adopting the following technical solution:

[0008] A method for quantitatively detecting blended coal components based on machine vision includes:

[0009] Obtain a blended coal image of the blended coal to be detected, where the blended coal to be detected includes multiple single coals;

[0010] Input the blended coal image into a pre-trained quantitative identification model for blended coal to obtain the quantitative detection result of the to-be-detected blended coal. The quantitative detection result includes the coal types and contents of each single coal in the to-be-detected blended coal.

[0011] Among them, the quantitative identification model for blended coal is trained by multiple target blended coal sample images, and each target blended coal sample image corresponds to the coal type and content of each single coal among multiple single coals.

[0012] The beneficial effects of the present invention are as follows: rapid and accurate quantitative detection of the components of blended coal is realized. Specifically, through image acquisition and a machine vision model, the identification and quantitative analysis of each single coal component in blended coal can be completed in a short time, significantly shortening the time compared with traditional chemical detection methods. Expensive chemical reagents and complex manual operations are avoided, the detection cost is reduced, and the economic benefit is improved.

[0013] On the basis of the above technical solution, the present invention can also be improved as follows.

[0014] Further, the construction method of the quantitative identification model for blended coal includes:

[0015] Step S21, construct a blended coal training set. The blended coal training set includes multiple target blended coal sample images. Each target blended coal sample image corresponds to a labeling result, and each labeling result identifies the coal type and content of each single coal among multiple single coals in the corresponding target blended coal sample image.

[0016] Step S22, for any one of the target blended coal sample images in the blended coal training set, input the target blended coal sample image into the U-net semantic segmentation model to obtain the target region and coal type corresponding to each single coal in the target blended coal sample image.

[0017] Step S23, for each target blended coal sample image, based on the target region and coal type corresponding to each single coal in the target blended coal sample image, calculate the area ratio of the target region corresponding to each single coal to the area of the target blended coal sample image. Based on the area ratio corresponding to each single coal, obtain the content of each single coal in the target blended coal sample image.

[0018] Step S24, based on the coal type and content of each single coal among multiple single coals in all target blended coal sample images, calculate the loss function, and update the model parameters of the U-net semantic segmentation model through the backpropagation algorithm. Repeat steps S22 to S24 until the number of iterations reaches the set number of iterations, and use the updated U-net semantic segmentation model as the quantitative identification model for blended coal.

[0019] The beneficial effects of adopting the above further solution are as follows: By identifying the coal types and contents of each single coal, the diversity and accuracy of the training data are ensured. The U-net semantic segmentation model is used to accurately segment the blended coal sample image, achieving precise positioning of the regions of each single coal component in the blended coal. Based on the target regions and coal types corresponding to each single coal in the target blended coal sample image, the area ratio of the target region corresponding to each single coal in the target blended coal sample image is calculated. Based on the area ratio corresponding to each single coal, the content of each single coal in the target blended coal sample image is obtained, improving the reliability of the detection result. By continuously iteratively updating the model parameters of the U-net semantic segmentation model, a blended coal quantitative recognition model with excellent performance is finally obtained, ensuring the stability and robustness of the model in practical applications.

[0020] Further, the construction of the blended coal training set includes:

[0021] Obtain the initial blended coal sample images of multiple target blended coal samples;

[0022] Input each of the initial blended coal sample images into a pre-trained single coal recognition model respectively, and perform single coal recognition on each of the initial blended coal sample images to obtain single coal recognition results, where the single coal recognition results include the coal type and position region of each single coal in the multiple single coals corresponding to each initial blended coal sample image;

[0023] Based on the single coal recognition results, use a set image annotation tool to label each of the initial blended coal sample images to obtain multiple target blended coal sample images, and use the multiple target blended coal sample images as the blended coal training set.

[0024] The beneficial effects of adopting the above further solution are as follows: Using the pre-trained single coal recognition model to perform single coal recognition on each initial blended coal sample image, the specific positions and type information of each single coal in each initial blended coal sample image are obtained, providing an accurate basis for subsequent image annotation and making the image annotation more accurate. Using a set image annotation tool to label each initial blended coal sample image, the generated target blended coal sample images not only contain rich image features but also carry clear single coal component region labels. These high-quality data sets contribute to the training and optimization of the U-net semantic segmentation model, thereby improving the accuracy and reliability of the final quantitative detection of blended coal components.

[0025] Further, the training method of the single coal recognition model includes:

[0026] Obtain the single coal images of multiple single coal samples;

[0027] Extract features from each of the single coal images respectively to obtain the feature information of each of the single coal images;

[0028] Classify the feature information of all the single - coal images based on the K - means clustering algorithm to obtain a clustering result. The clustering result includes multiple clusters, and each cluster represents the feature information of a coal type.

[0029] Based on the clustering result, use an image annotation tool to identify the coal types for each of the single - coal images respectively.

[0030] Construct a single - coal training set based on multiple single - coal images and the coal - type identification corresponding to each single - coal image.

[0031] Train the yolov5 object classification algorithm based on the single - coal training set to obtain a single - coal recognition model.

[0032] The beneficial effect of adopting the above - mentioned further solution is as follows: It can classify and identify single - coal images efficiently and accurately. First, obtain the single - coal images of multiple single - coal samples and extract their features, including gray - scale features, morphological features, and contour features, ensuring the diversity and richness of the data. Then, use the K - means clustering algorithm to classify the feature information of all single - coal images, generating multiple clusters, and each cluster represents the feature information of a coal type, thus achieving effective differentiation of different coal types. Subsequently, based on the clustering result, use an image annotation tool to identify the coal types for each single - coal image, forming a high - quality single - coal training set. Finally, use the yolov5 object classification algorithm to train the single - coal training set to obtain an efficient single - coal recognition model. This model not only improves the accuracy of single - coal recognition but also significantly shortens the recognition time, providing a solid foundation for the subsequent quantitative detection of blended - coal components.

[0033] Further, each of the single - coal samples and each of the blended - coal samples are obtained through the crushing and screening sample - preparation method, and each blended - coal sample includes components of no more than three single - coal samples.

[0034] The beneficial effect of adopting the above - mentioned further solution is as follows: It can ensure the standardization of the preparation process of blended - coal samples, improve the consistency and repeatability of the samples. Each blended - coal sample contains components of no more than three single - coal samples, which helps to simplify the complexity of image processing and model training, improve the recognition accuracy and efficiency. At the same time, the crushing and screening sample - preparation method ensures the uniformity and representativeness of the samples, enhancing the adaptability of the model to different blended - coal combinations.

[0035] Further, when extracting the features of each single - coal image respectively to obtain the feature information of each single - coal image, it includes:

[0036] Calculate the overall gray - scale value of each single - coal image to obtain the gray - scale feature of each single - coal image.

[0037] Based on a preset mathematical operation shape method, morphological features of each of the single coal images are extracted to obtain the morphological features of each of the single coal images;

[0038] The principal component analysis method is used to extract the texture features of each of the single coal images to obtain the contour features of each of the single coal images;

[0039] Based on the gray-scale features, the morphological features, and the contour features of each of the single coal images, the feature information of each of the single coal images is determined.

[0040] The beneficial effect of adopting the above further solution is that it can effectively improve the accuracy of feature extraction of single coal images.

[0041] In a second aspect, the present application provides an application of a method for quantitatively detecting mixed coal components based on machine vision, and the following technical solution is adopted:

[0042] An application of a method for quantitatively detecting mixed coal components based on machine vision in coal quality inspection.

[0043] In a third aspect, the present application provides a device for quantitatively detecting mixed coal components based on machine vision, and the following technical solution is adopted:

[0044] A device for quantitatively detecting mixed coal components based on machine vision, comprising:

[0045] An acquisition module for acquiring a mixed coal image of the mixed coal to be detected, wherein the mixed coal to be detected includes multiple single coal components;

[0046] An identification module for inputting the mixed coal image into a pre-trained mixed coal quantitative identification model to obtain a quantitative detection result of the mixed coal to be detected, where the quantitative detection result includes the coal types and contents of each of the single coals in the mixed coal to be detected;

[0047] Wherein, the mixed coal quantitative identification model is trained by multiple target mixed coal sample images, and each of the target mixed coal sample images corresponds to the coal type and content of each of the single coals in multiple single coals.

[0048] In a fourth aspect, the present application provides an electronic device, and the following technical solution is adopted:

[0049] An electronic device includes a memory and a processor, and a computer program capable of being loaded and executed by the processor for the method for quantitatively detecting mixed coal components based on machine vision according to any one of the first aspects is stored on the memory.

[0050] In a fifth aspect, the present application provides a computer-readable storage medium, and the following technical solution is adopted:

[0051] A computer-readable storage medium stores a computer program that can be loaded and executed by a processor to perform the machine vision-based quantitative detection method for blended coal components described in any one of the first aspects.

[0052] Additional aspects and advantages of the present application will be given in part in the following description, which will become apparent from the following description, or can be learned through the practice of the present application.

[0053] The present invention has the following beneficial effects:

[0054] 1. The present invention mainly based on the machine vision model and vision processing technology realizes the rapid quantitative detection of blended coal. The time to analyze and fit the blending ratio of blended coal components by the image method can reach the second level.

[0055] 2. The object detected by the present invention can be blended coal particles at the micron scale under a microscope lens or blended coal particles at the millimeter level, and the applicability of the detection sample is wide. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic flow chart of the machine vision-based quantitative detection method for blended coal components provided by an embodiment of the present invention;

[0057] Figure 2 It is a schematic diagram of the training of the blended coal quantitative recognition model provided by an embodiment of the present invention;

[0058] Figure 3 It is a schematic diagram of a single coal sample image provided by an embodiment of the present invention;

[0059] Figure 4 It is a schematic diagram of a blended coal sample image provided by an embodiment of the present invention;

[0060] Figure 5 It is a schematic diagram of the recognition result of a single coal provided by an embodiment of the present invention;

[0061] Figure 6 It is a schematic diagram of the quantitative detection result provided by an embodiment of the present invention;

[0062] Figure 7 It is a schematic structural diagram of the machine vision-based quantitative detection device for blended coal components provided by an embodiment of the present invention;

[0063] Figure 8 It is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the following will clearly and completely describe the technical solutions in the embodiments of this application with reference to the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are some, but not all, of the embodiments of this application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of this application without creative efforts shall fall within the scope of protection of this application.

[0065] In addition, the term "and / or" in this document is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, unless otherwise specified.

[0066] As Figure 1 shown, the embodiments of this application provide a method for quantitatively detecting the components of blended coal based on machine vision, the application of the method for quantitatively detecting the components of blended coal based on machine vision in coal quality inspection. This method can be executed by an electronic device, which can be a server or a mobile terminal device. The server can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services; the mobile terminal device can be a laptop, a desktop computer, etc., but is not limited thereto.

[0067] This method mainly includes (Steps S1 to S2):

[0068] Step S1: Obtain a blended coal image of the blended coal to be detected, where the blended coal to be detected includes multiple single coal components;

[0069] Step S2: Input the blended coal image into a pre-trained blended coal quantitative recognition model to obtain a quantitative detection result of the blended coal to be detected. The quantitative detection result includes the coal types and contents of each single coal in the blended coal to be detected;

[0070] Among them, the blended coal quantitative recognition model is trained by multiple target blended coal sample images, and each target blended coal sample image corresponds to the coal type and content of each single coal in multiple single coals.

[0071] In the embodiments of the present application, first, coal types are selected and mixed. The commonly used coal types for power generation in thermal power plants are selected and mixed in a certain proportion to obtain blended coal. For example, 20 common coal types in coal-fired power plants can be selected and mixed in different proportions, and each portion of blended coal contains no more than three components of coal. As many coal type categories as possible are covered to improve the generalization ability of the model. Specifically, the following several coal types can be selected: anthracite, lean coal, meager coal, coking coal, fat coal, gas coal, lignite, etc. Since the properties of each coal type are different, through diverse coal type selection, the influence of a single coal type on the model can be reduced to a certain extent, making the model more robust.

[0072] In the embodiments of the present application, before constructing the quantitative identification model of blended coal, first, 5 kinds of coal commonly used for power generation in thermal power plants are adopted. According to their sources, they are respectively named HSQ, JSB2, TC, WCW, and YH. The 5 kinds of coal are blended in pairs at a ratio of 1:1 to obtain 10 kinds of blended coal;

[0073] Secondly, according to the preparation method of coal petrography analysis samples, the blended coal samples are ground, broken, inlaid, and polished. This is to ensure that the surface of the samples is flat, facilitating the image acquisition device to obtain high-quality images.

[0074] After that, images of the blended coal samples and single coal samples are collected. In the embodiments of the present application, an image acquisition device is used to collect images of single coal samples and blended coal samples with known mixing ratios, and image features are extracted. The image acquisition device can be a high-resolution microscope or an industrial camera, equipped with appropriate lenses and light sources to ensure high clarity of the collected images. Specifically, an optical microscope can be used with a digital camera to take microscopic images of single coal samples and blended coal samples. There are no less than 50 images for each single coal sample, and no less than 50 images for each blended coal sample type. They are used for model training and verification.

[0075] Optionally, as Figure 2 shown, the method for constructing the quantitative identification model of blended coal includes:

[0076] Step S21, constructing a blended coal training set, where the blended coal training set includes multiple target blended coal sample images, each of the target blended coal sample images corresponds to a labeling result, and each labeling result identifies the coal type and content of each single coal in the corresponding target blended coal sample image;

[0077] In the embodiments of the present application, as Figure 3 and Figure 4 shown, taking HSQ and TC coal as an example, images of HSQ and TQ single coal samples are selected, with no less than 50 images for each single coal type. They are used as the training set to train the recognition model for predicting the image mixing ratio of HSQ:TC = 1:1 blended coal.

[0078] In this embodiment, constructing the blended coal training set includes the following sub-steps:

[0079] Step S211: Obtain initial blended coal sample images of multiple target blended coal samples;

[0080] Step S212: Input each of the initial blended coal sample images into a pre-trained single coal recognition model respectively, perform single coal recognition on each of the initial blended coal sample images, and obtain single coal recognition results. The single coal recognition results include the coal type and position area of each single coal in the multiple single coals corresponding to each initial blended coal sample image;

[0081] In the embodiment of the present application, the yolov5 algorithm is selected to construct the single coal recognition model, using the feature vector of the coal petrographic microconstituent image as the input and the type of coal petrology as the output. The recognition effect is as Figure 5 shown.

[0082] Specifically, the training method of the single coal recognition model includes:

[0083] Obtain single coal images of multiple single coal samples;

[0084] Extract features from each of the single coal images respectively to obtain the feature information of each single coal image;

[0085] Classify the feature information of all the single coal images based on the K-means clustering algorithm to obtain a clustering result. The clustering result includes multiple clusters, and each cluster represents the feature information of a coal type;

[0086] Based on the clustering result, use an image annotation tool to label the coal types of each of the single coal images respectively;

[0087] Construct a single coal training set based on multiple single coal images and the coal type labels corresponding to each single coal image;

[0088] Train the yolov5 object classification algorithm based on the single coal training set to obtain a single coal recognition model.

[0089] Using the pre-trained single coal recognition model to perform single coal recognition on each blended coal sample image, the specific positions and type information of each single coal component in each blended coal sample image are obtained, providing an accurate basis for subsequent image annotation and making the labeling of blended coal sample images more accurate.

[0090] Step S213: Based on the single coal recognition results, use a set image annotation tool to label each of the initial blended coal sample images to obtain multiple target blended coal sample images, and use the multiple target blended coal sample images as the blended coal training set.

[0091] In the embodiments of the present application, specifically, the step of separately extracting features from each of the single coal images to obtain the feature information of each of the single coal images includes:

[0092] Calculating the overall gray value of each of the single coal images to obtain the gray feature of each of the single coal images; the gray scale of the coal sample is 8 bits

[0093] Based on a preset mathematical operation shape method, performing morphological feature extraction on each of the single coal images to obtain the morphological features of each of the single coal images. The mathematical operation shape method includes Fourier descriptors, wavelet descriptors, etc. The morphological features include morphological indexes such as convexity, roundness, squareness, contour irregularity, contour kurtosis, and eccentricity;

[0094] Using the principal component analysis method to extract the texture features of each of the single coal images to obtain the contour features of each of the single coal images;

[0095] Based on the gray feature of each of the single coal images, the morphological feature of each of the single coal images, and the contour feature of each of the single coal images, determining the feature information of each of the single coal images. It can effectively improve the accuracy of feature extraction of single coal images.

[0096] In the embodiments of the present application, the labelme software, a marking tool, is selected to identify the coal types of each of the single coal images separately.

[0097] Step S22: For any one of the target blended coal sample images in the blended coal training set, input the target blended coal sample image into the U-net semantic segmentation model to obtain the target regions and coal types corresponding to each single coal in the target blended coal sample image;

[0098] Step S23: For each of the target blended coal sample images, based on the target regions and coal types corresponding to each single coal in the target blended coal sample image, calculate the area ratio of the target region corresponding to each single coal to the target blended coal sample image. Based on the area ratio corresponding to each single coal, obtain the content of each single coal in the target blended coal sample image;

[0099] Step S24: Based on the coal types and contents of each single coal in multiple single coals in all target blended coal sample images, calculate the loss function, and update the model parameters of the U-net semantic segmentation model through the backpropagation algorithm. Repeat steps S22 to S24 until the number of iterations reaches the set number of iterations, and use the updated U-net semantic segmentation model as the blended coal quantitative recognition model.

[0100] In this embodiment, as Figure 6As shown, through the U-net semantic segmentation model, the image features of different components in blended coal are learned at the pixel level. The trained model can transform the pixel regions occupied by the coal sample pictures of different components into different colors, and calculate the area ratio of the pixel regions occupied by the pictures of different coal quality components.

[0101] Through iterative calculation, the model is used to calculate the ratio of the pixel areas occupied by each single coal component in 50 blended coal pictures with HSQ:TC = 1:1, find the average value and evaluate whether it is close to the pre-prepared ratio; the prediction given by the model is HSQ:TC = 1.11, which is close to the actual value of HSQ:TC = 1:1.

[0102] By identifying multiple single coal species regions of different colors, the diversity and accuracy of the training data are ensured. The U-net semantic segmentation model is used to accurately segment the blended coal sample image, realizing the precise positioning of each single coal component region in the blended coal. Based on the segmentation results, the area ratio of different single coal component regions occupying the target blended coal sample image is calculated, obtaining the content information of each single coal component and improving the reliability of the detection results. By continuously iteratively updating the model parameters of the U-net semantic segmentation model, a blended coal quantitative recognition model with excellent performance is finally obtained, ensuring the stability and robustness of the model in practical applications.

[0103] This method can complete the identification and quantitative analysis of each single coal component in blended coal in a short time through image acquisition and machine vision models, significantly shortening the time compared with traditional chemical detection methods. Using the yolov5 algorithm and the U-net semantic segmentation model, each single coal component in the blended coal image can be accurately extracted, and its proportion in the image can be calculated, so as to accurately predict the true proportion of each component in the blended coal. It avoids expensive chemical reagents and complex manual operations, reduces the detection cost, and improves the economic benefits.

[0104] Figure 7 This is a schematic structural diagram of a blended coal component quantitative detection device 200 based on machine vision provided by an embodiment of the present invention.

[0105] As Figure 7 shown, a blended coal component quantitative detection method device 200 based on machine vision mainly includes:

[0106] An acquisition module 201, configured to acquire a blended coal image of the blended coal to be detected, where the blended coal to be detected includes multiple single coal components;

[0107] An identification module 202, configured to input the blended coal image into a pre-trained blended coal quantitative recognition model to obtain a quantitative detection result of the blended coal to be detected, where the quantitative detection result includes the coal types and contents of each of the single coals in the blended coal to be detected;

[0108] Among them, the blended coal quantitative identification model is obtained by training multiple target blended coal sample images, and each of the target blended coal sample images corresponds to the coal types and contents of each single coal in multiple single coals.

[0109] In one example, the modules in any of the above devices can be one or more integrated circuits configured to implement the above methods. For example: one or more application specific integrated circuits (ASICs), or, one or more digital signal processors (DSPs), or, one or more field programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0110] Again, when the modules in the device can be implemented in the form of a processing element scheduler, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call programs. Again, these modules can be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0111] In this application, names may be assigned to various objects such as various messages / information / devices / network elements / systems / devices / actions / operations / processes / concepts, etc. It can be understood that these specific names do not constitute a limitation on the relevant objects, and the assigned names may change with factors such as scenarios, contexts, or usage habits. The understanding of the technical meaning of the technical terms in this application should mainly be determined from the functions and technical effects reflected / executed in the technical solutions.

[0112] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices, and modules described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0113] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but this implementation should not be considered to exceed the scope of this application.

[0114] Figure 8This is a structural block diagram of an electronic device 300 according to an embodiment of the present application.

[0115] As Figure 8 shown, the electronic device 300 includes a processor 301 and a memory 302, and may further include one or more of an information input / output (I / O) interface 303, a communication component 304, and a communication bus 305.

[0116] Among them, the processor 301 is used to control the overall operation of the electronic device 300 to complete all or part of the steps in the above-mentioned machine vision-based mixed coal component quantitative detection method; the memory 302 is used to store various types of data to support the operation of the electronic device 300. These data may include, for example, instructions for any application or method operating on the electronic device 300, and application-related data. The memory 302 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disc, or one or more of them.

[0117] The I / O interface 303 provides an interface between the processor 301 and other interface modules. The above other interface modules may be a keyboard, a mouse, buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 304 is used to test the wired or wireless communication between the electronic device 300 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G, or 4G, or a combination of one or more of them. Therefore, the corresponding communication component 304 may include: a Wi-Fi component, a Bluetooth component, and an NFC component.

[0118] The communication bus 305 may include a path for transmitting information among the above components. The communication bus 305 may be a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The communication bus 305 may be divided into an address bus, a data bus, a control bus, etc.

[0119] The electronic device 300 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components, and is used to execute the method for quantitatively detecting the blending coal components based on machine vision given in the above embodiments.

[0120] The computer-readable storage medium provided by the embodiments of the present application will be introduced below. The computer-readable storage medium described below can be correspondingly referred to the method for quantitatively detecting the blending coal components based on machine vision described above.

[0121] The present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method for quantitatively detecting the blending coal components based on machine vision are implemented.

[0122] The computer-readable storage medium may include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

[0123] The term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0124] The above description is only a preferred embodiment of the present application and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the application involved in the present application is not limited to the technical solutions formed by the specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the foregoing application concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions applied in the present application.

Claims

1. A method for quantitatively detecting the components of blended coal based on machine vision, characterized in that, Including: Obtain a blended coal image of the blended coal to be detected, where the blended coal to be detected includes multiple single coals; Input the blended coal image into a pre-trained blended coal quantitative recognition model to obtain a quantitative detection result of the blended coal to be detected, where the quantitative detection result includes the coal type and content of each single coal in the blended coal to be detected; Among them, the blended coal quantitative recognition model is trained by multiple target blended coal sample images, and each target blended coal sample image corresponds to the coal type and content of each single coal in multiple single coals.

2. The quantitative detection method for blended coal components based on machine vision according to claim 1, wherein, The construction method of the blended coal quantitative recognition model includes: Step S21, construct a blended coal training set, where the blended coal training set includes multiple target blended coal sample images, each target blended coal sample image corresponds to a labeling result, and each labeling result identifies the coal type and content of each single coal in the multiple single coals in the corresponding target blended coal sample image; Step S22, for any one of the target blended coal sample images in the blended coal training set, input the target blended coal sample image into the U-net semantic segmentation model to obtain the target area and coal type corresponding to each single coal in the target blended coal sample image; Step S23, for each target blended coal sample image, based on the target area and coal type corresponding to each single coal in the target blended coal sample image, calculate the area ratio of the target area corresponding to each single coal to the target blended coal sample image, and based on the area ratio corresponding to each single coal, obtain the content of each single coal in the target blended coal sample image; Step S24, based on the coal type and content of each single coal in multiple single coals in all target blended coal sample images, calculate the loss function, and update the model parameters of the U-net semantic segmentation model through the backpropagation algorithm. Repeat steps S22 to S24 until the number of iterations reaches the set number of iterations, and use the updated U-net semantic segmentation model as the blended coal quantitative recognition model.

3. The method for quantitatively detecting the mixed coal components based on machine vision according to claim 2, wherein, The construction of the blended coal training set includes: Obtain initial blended coal sample images of multiple target blended coals; Input each initial blended coal sample image into a pre-trained single coal recognition model respectively, and perform single coal recognition on each initial blended coal sample image to obtain a single coal recognition result, where the single coal recognition result includes the coal type and position area of each single coal in the multiple single coals corresponding to each initial blended coal sample image; Based on the single coal recognition result, use a set image annotation tool to label each initial blended coal sample image to obtain multiple target blended coal sample images, and use the multiple target blended coal sample images as the blended coal training set.

4. A method for quantitatively detecting the components of blended coal based on machine vision according to claim 3, characterized in that, The training method of the single coal recognition model includes: Obtain single coal images of multiple single coals; Extract features from each single coal image respectively to obtain the feature information of each single coal image; Classify the feature information of all single coal images based on the K-means clustering algorithm to obtain a clustering result, where the clustering result includes multiple clusters, and each cluster represents the feature information of a coal type; Based on the clustering result, use an image annotation tool to label the coal type of each single coal image respectively; Construct a single coal training set based on multiple single coal images and the coal type identification corresponding to each of the single coal images; Train the yolov5 object classification algorithm based on the single coal training set to obtain a single coal recognition model.

5. The quantitative detection method for blended coal components based on machine vision according to claim 4, wherein, Each of the single coal samples and each of the blended coal samples are obtained by the crushing, screening and sample preparation method, and each of the blended coal samples includes components of no more than three single coal samples.

6. The quantitative detection method for blended coal components based on machine vision according to claim 4, wherein Respectively extract features from each of the single coal images to respectively obtain the feature information of each of the single coal images, including: Calculate the overall gray value of each of the single coal images to obtain the gray feature of each of the single coal images; Based on a preset mathematical operation shape method, extract the morphological features of each of the single coal images to obtain the morphological features of each of the single coal images; Adopt the principal component analysis method to extract the texture features of each of the single coal images to obtain the contour features of each of the single coal images; Based on the gray feature of each of the single coal images, the morphological feature of each of the single coal images and the contour feature of each of the single coal images, determine the feature information of each of the single coal images.

7. Application of a method for quantitatively detecting the components of blended coal based on machine vision according to any one of claims 1-6 in coal quality inspection.

8. A device for quantitatively detecting the components of blended coal based on machine vision, characterized in that, Including: An acquisition module for acquiring a blended coal image of the blended coal to be detected, where the blended coal to be detected includes multiple single coal components; An identification module for inputting the blended coal image into a pre-trained blended coal quantitative identification model to obtain a quantitative detection result of the blended coal to be detected, where the quantitative detection result includes the coal type and content of each of the single coals in the blended coal to be detected; Wherein, the blended coal quantitative identification model is trained by multiple target blended coal sample images, and each of the target blended coal sample images corresponds to the coal type and content of each of the single coals in multiple single coals.

9. An electronic device, characterized in that, Including a processor, and the processor is coupled to a memory; The processor is configured to execute a computer program stored in the memory so that the electronic device executes the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, Including a computer program or instruction, when the computer program or instruction runs on a computer, the computer is caused to execute the method according to any one of claims 1-6.

Images