Automatic unloading and cleaning system and method for frozen coal in train compartment

Through the automatic unloading and cleaning system of frozen coal in the train compartment, visual guidance and intelligent algorithms are used to automatically identify the state of frozen coal, and the vibration frequency of crusher is adjusted, which solves the problem of low unloading efficiency of frozen coal, realizes automatic unloading and cleaning, and reduces labor intensity and time.

CN118529515BActive Publication Date: 2024-11-05SHANDONG HAIDA ROBOT TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410750028.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-12
Publication Date
2024-11-05
Estimated Expiration
2044-06-12

AI Technical Summary

Technical Problem

During the transportation of coal from railway trains, frozen coal materials make it impossible to unload and clean the coal in the carriage smoothly. The existing artificial ice breaking method is labor-intensive and inefficient.

Method used

An automatic unloading and cleaning system for frozen coal in the train car is designed, using rails, mobile gantry platform, lifting mechanism, crusher and visual guidance system. The state of frozen coal is automatically identified through image processing and intelligent algorithms, and the vibration frequency of the crusher is adjusted to crush Frozen coal and clean and recycling.

Benefits of technology

It reduces the labor intensity of the workers, improves the coal unloading speed, automatically recognizes and crushes frozen coal, and avoids the subjective problem of artificial preset vibration frequency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118529515B_ABST
    Figure CN118529515B_ABST
Patent Text Reader

Abstract

A system and method for automatically unloading and cleaning frozen coal in a train carriage, comprising: rails arranged along the travel direction of a railway vehicle; a mobile gantry platform movably mounted on the rails; a lifting mechanism installed on the mobile gantry platform, wherein the lifting mechanism includes a lifting mechanism for a frozen coal crushing system and a lifting mechanism for a cleaning and recovery system; a frozen coal crushing system is installed in the lifting mechanism for the frozen coal crushing system, wherein the frozen coal crushing system includes a crusher and a visual guidance system, a control system is arranged on the mobile gantry platform, the visual guidance system is used to identify the state information of the frozen coal in the vehicle, and the control system is used to control the crusher. The frozen coal image collected by the visual guidance system is used to learn and capture the multi-scale feature representation of the coal state of the frozen coal image, so as to intelligently decode and restore the recommended value of the initial vibration frequency based on this multi-scale feature representation of the frozen coal state.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of intelligent unloading and cleaning, and more specifically, to an automatic unloading and cleaning system for frozen coal in a train compartment and a method thereof. Background Art

[0002] Railway train coal transportation has the advantages of high transportation efficiency, low cost of long-distance and large-volume transportation, energy saving and environmental protection. In order to prevent coal dust from spreading around during long-distance transportation and loading, it is usually necessary to spray water on the coal to suppress dust. However, in winter, due to the cold weather, freezing of coal during train transportation is a common and difficult problem to solve. In particular, the coal at the bottom of the train and close to the carriage freezes during loading and unloading, making it impossible to unload and clean the coal inside the carriage smoothly.

[0003] In order to ensure that the coal inside the carriage can be unloaded smoothly and to complete the cleaning of the residual materials in the carriage, manual ice breaking is required. For example, the operator uses tools such as steel chisels to break the frozen coal. However, this manual ice breaking method is labor-intensive and has a harsh working environment, which seriously affects the speed of coal unloading and cleaning.

[0004] Therefore, an optimized automatic unloading and cleaning system for frozen coal in train compartments and a method thereof are desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The present application provides a system and method for automatically unloading and cleaning frozen coal in a train compartment, which includes: arranging steel rails along the direction of travel of the rails; movably installing a mobile gantry platform on the steel rails; installing a lifting mechanism on the mobile gantry platform, wherein the lifting mechanism includes a platform lifting mechanism and a cleaning and recovery system lifting mechanism; installing a cleaning and recovery system on the cleaning and recovery system lifting mechanism, wherein the cleaning and recovery system includes a crusher, a visual guidance system and a control system, using the visual guidance system to identify the state information of the frozen coal in the vehicle, and using the control system to control the crusher. In this way, the frozen coal image collected by the visual guidance system is combined with an intelligent algorithm to perform multi-scale feature mining and fusion on it, so as to learn and capture the multi-scale feature representation of the coal state of the frozen coal image, so as to intelligently decode and restore the recommended value of the initial vibration frequency based on this multi-scale frozen coal state feature representation.

[0006] In a first aspect, a system for automatically unloading and cleaning frozen coal from a train compartment is provided, comprising:

[0007] Steel rails set along the direction of travel of railway vehicles;

[0008] a mobile gantry platform movably mounted on the rails;

[0009] A lifting mechanism installed on the mobile gantry platform, wherein the lifting mechanism includes a frozen coal crushing system lifting mechanism and a cleaning and recovery system lifting mechanism; and

[0010] A frozen coal crushing system installed on the lifting mechanism of the frozen coal crushing system, wherein the frozen coal crushing system includes a crusher and a visual guidance system. A control system is arranged on the mobile gantry platform. The visual guidance system is used to identify the status information of the frozen coal in the vehicle, and the control system is used to control the crusher.

[0011] In the above-mentioned automatic unloading and cleaning system for frozen coal in train carriages, the mobile gantry platform adopts an assembled frame structure, walking mechanisms are respectively arranged on both sides of the bottom of the mobile gantry platform, the steel rails are laid on the bottom of the mobile gantry platform, and the walking mechanism adopts a synchronous control method.

[0012] In the above-mentioned automatic unloading and cleaning system for frozen coal in train carriages, the crusher is driven by a hydraulic motor for vibration excitation, wherein the crusher adjusts the rotation speed of the hydraulic motor in real time to adjust the vibration frequency to reach the resonance frequency of the frozen coal to complete the crushing of the frozen coal.

[0013] In the above-mentioned automatic unloading and cleaning system for frozen coal in a train compartment, the control system also includes an initial vibration frequency controller; wherein the initial vibration frequency controller includes: a frozen coal image acquisition module, used to acquire the frozen coal image collected by the visual guidance system; a multi-scale feature extraction module, used to perform multi-scale feature extraction on the frozen coal image to obtain a multi-scale feature map of the coal state; a local feature saliency module, used to pass the multi-scale feature map of the coal state through a local feature saliency device based on a convolution kernel attention mechanism to obtain a salient multi-scale feature map of the coal state; and a recommended value determination module for the initial vibration frequency, used to determine the recommended value of the initial vibration frequency based on the salient multi-scale feature map of the coal state.

[0014] In the above-mentioned automatic unloading and cleaning system for frozen coal in a train compartment, the multi-scale feature extraction module includes: shallow and deep feature extraction units, which are used to perform shallow feature extraction and deep feature extraction on the frozen coal image respectively to obtain a shallow feature map of the frozen coal state and a deep feature map of the frozen coal state; and a feature fusion attention guidance unit, which is used to pass the shallow feature map of the frozen coal state and the deep feature map of the frozen coal state through a feature fusion attention guidance network to obtain the multi-scale feature map of the coal state.

[0015] In the above-mentioned automatic unloading and cleaning system for frozen coal in a train compartment, the shallow and deep feature extraction units are used to: pass the frozen coal image through a frozen coal shallow feature extractor based on a first convolutional neural network model to obtain a shallow feature map of the frozen coal state; and pass the frozen coal state shallow feature map through a frozen coal deep feature extractor based on a second convolutional neural network model to obtain a deep feature map of the frozen coal state.

[0016] In the above-mentioned automatic unloading and cleaning system for frozen coal in a train compartment, the feature fusion attention guiding unit is used to: perform global average pooling along the channel dimension on the deep feature map of the frozen coal state to obtain an attention feature vector; pass the attention feature vector through a fully connected layer to obtain an attention encoding feature vector; perform weighted multiplication on the shallow feature map of the frozen coal state using each eigenvalue in the attention encoding feature vector as a weight to obtain an attention adjustment feature map; and perform positional addition processing on the attention adjustment feature map and the deep feature map of the frozen coal state to obtain the multi-scale feature map of the coal state.

[0017] In the above-mentioned automatic unloading and cleaning system for frozen coal in train compartments, the local feature significant module includes: a nonlinear convolution operation unit, which is used to perform a nonlinear convolution operation on the multi-scale feature map of the coal state through a multi-scale convolution kernel group to obtain a first nonlinear convolution feature map, a second nonlinear convolution feature map, a third nonlinear convolution feature map and a fourth nonlinear convolution feature map; a global representation processing unit, which is used to process the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map and the fourth nonlinear convolution feature map with the following global representation formula to obtain a global representation feature vector; wherein the global representation formula is:

[0018] ;

[0019] ;

[0020] in, , , and are respectively the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map and the fourth nonlinear convolution feature map, is a multi-scale nonlinear convolution fusion feature map, Indicates cascade processing, represents the first eigenvalues, Represents the multi-scale nonlinear convolution fusion feature map The feature matrix is ​​processed by global mean pooling. and They are respectively the first The height and width of the feature matrix, is the first In the feature matrix The invention relates to a method for processing the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map, the fourth nonlinear convolution feature map and the concentrated global characterization feature vector according to the following attention weight application formula to obtain the significant coal state multi-scale feature map; wherein the attention weight application formula is:

[0021] ;

[0022] ;

[0023] ;

[0024] ;

[0025] ;

[0026] ;

[0027] in, , , , Respectively represent the first nonlinear convolution feature map , the second nonlinear convolution feature map , the third nonlinear convolution feature map And the fourth nonlinear convolution feature map The weighted probability vector, represents the condensed global representation feature vector corresponding to the first nonlinear convolution feature map The vector elements of the channel, , and Then, the condensed global representation feature vectors corresponding to the second nonlinear convolution feature map, the third nonlinear convolution feature map, and the fourth nonlinear convolution feature map are represented respectively. The vector elements of the channel, is the significant coal state multi-scale characteristic map, Indicates cascade processing.

[0028] In the above-mentioned automatic unloading and cleaning system for frozen coal in a train compartment, the recommended value determination module for the initial vibration frequency is used to: pass the significant coal state multi-scale feature map through a decoder-based initial vibration frequency recommender to obtain a decoded value, and the decoded value is used to represent the recommended value of the initial vibration frequency.

[0029] In a second aspect, a method for automatically unloading and cleaning frozen coal from a train compartment is provided, comprising:

[0030] The rails are arranged along the direction of travel of the railway vehicles;

[0031] A movable gantry platform is movably installed on the steel rail;

[0032] A lifting mechanism is installed on the mobile gantry platform, wherein the lifting mechanism includes a frozen coal crushing system lifting mechanism and a cleaning and recovery system lifting mechanism;

[0033] The frozen coal crushing system is installed on the lifting mechanism of the frozen coal crushing system, wherein the frozen coal crushing system includes a crusher and a visual guidance system, a control system is arranged on the mobile gantry platform, the visual guidance system is used to identify the state information of the frozen coal in the vehicle, and the control system is used to control the crusher;

[0034] Among them, the control system also includes an initial vibration frequency controller, which is used to obtain a frozen coal image collected by the visual guidance system; multi-scale feature extraction is performed on the frozen coal image to obtain a multi-scale feature map of coal state; the multi-scale feature map of coal state is passed through a local feature salient device based on a convolution kernel attention mechanism to obtain a significant multi-scale feature map of coal state; and a recommended value of the initial vibration frequency is determined based on the significant multi-scale feature map of coal state.

[0035] Compared with the prior art, the automatic unloading and cleaning system and method for frozen coal in a train compartment provided by the present application include: setting rails along the travel direction of the railway vehicle; movably installing a mobile gantry platform on the rails; installing a lifting mechanism on the mobile gantry platform, wherein the lifting mechanism includes a frozen coal crushing system lifting mechanism and a cleaning and recovery system lifting mechanism; installing a frozen coal crushing system on the frozen coal crushing system lifting mechanism, wherein the frozen coal crushing system includes a crusher and a visual guidance system, and setting a control system on the mobile gantry platform, using the visual guidance system to identify the state information of the frozen coal in the vehicle, and using the control system to control the crusher. In this way, the frozen coal image collected by the visual guidance system is combined with an intelligent algorithm to perform multi-scale feature mining and fusion on it, so as to learn and capture the multi-scale feature representation of the coal state of the frozen coal image, so as to intelligently decode and restore the recommended value of the initial vibration frequency based on this multi-scale frozen coal state feature representation. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0037] Figure 1 It is a structural schematic diagram of a system for automatically unloading and cleaning frozen coal from a train compartment according to an embodiment of the present application.

[0038] Figure 2 It is a block diagram of the initial vibration frequency controller in the automatic unloading and cleaning system for frozen coal in a train compartment according to an embodiment of the present application.

[0039] Figure 3 The present invention is a flow chart of a method for automatically unloading and cleaning frozen coal from a train compartment according to an embodiment of the present application.

[0040] Figure 4 This is an application scenario diagram of the automatic unloading and cleaning system for frozen coal in a train compartment according to an embodiment of the present application. DETAILED DESCRIPTION

[0041] The following will describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0042] Unless otherwise specified, all technical and scientific terms used in the embodiments of the present application have the same meaning as those commonly understood by those skilled in the art of the present application. The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit the scope of this application.

[0043] In the embodiments of the present application, it should be noted that, unless otherwise specified and limited, the term "connection" should be understood in a broad sense. For example, it can be an electrical connection or a connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0044] It should be noted that the terms "first\second\third" involved in the embodiments of the present application are only used to distinguish similar objects, and do not represent a specific order for the objects. It is understandable that the specific order or sequence of "first\second\third" can be interchanged where permitted. It should be understood that the objects distinguished by "first\second\third" can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0045] After introducing the basic principles of the present application, various non-limiting embodiments of the present application will be described in detail with reference to the accompanying drawings.

[0046] Railway train coal transportation is currently the main mode of coal transportation in my country and most countries in the world due to its high transportation efficiency, low cost of long-distance and large-scale transportation, and outstanding advantages in energy conservation and environmental protection. In winter, due to the cold weather, long-distance transportation and coal loading and transportation, water needs to be sprayed on the coal to prevent the spread of coal dust. The freezing of coal during train transportation is a common and difficult problem to solve. Due to the influence of the geographical environment in some areas, during the process of train transportation of coal, although the freezing of coal on the upper part of the train is not too serious, the coal under the car and close to the car is severely frozen during the loading and unloading process, making it impossible for the coal inside the car to be unloaded smoothly.

[0047] In order to ensure the smooth unloading of coal inside the carriage, it is necessary to manually break the ice, using tools such as steel chisels to break the frozen coal. The labor intensity of the operators is high and the working environment is poor, which seriously affects the unloading speed of the coal. Therefore, how to provide a train frozen coal breaking device and method is a technical problem that technicians in this field need to solve urgently.

[0048] The purpose of the present application is to provide an automatic unloading device for frozen coal in a train compartment, which replaces the manual work of thawing, unloading and cleaning the frozen coal in the train compartment, thereby reducing the labor intensity of the operators and increasing the unloading speed of the coal.

[0049] In one embodiment of the present application, Figure 1 Schematic diagram of the structure of the automatic unloading and cleaning system for frozen coal in a train compartment according to an embodiment of the present application. Figure 1As shown, according to an embodiment of the present application, the automatic unloading and cleaning system 100 for frozen coal in a train compartment includes: a steel rail 1 arranged along the traveling direction of the railway vehicle; a mobile gantry platform 2 movably mounted on the steel rail 1; a lifting mechanism 3 mounted on the mobile gantry platform 2, wherein the lifting mechanism 3 includes a frozen coal crushing system lifting mechanism 31 and a cleaning and recovery system lifting mechanism 32; and a frozen coal crushing system 4 mounted on the frozen coal crushing system lifting mechanism 31, wherein the frozen coal crushing system 4 includes a crusher 41 and a visual guidance system 42, and a control system 43 is arranged on the mobile gantry platform, wherein the visual guidance system 42 is used to identify the status information of the frozen coal in the vehicle, and the control system 43 is used to control the crusher 41.

[0050] In a specific embodiment of the present application, the mobile gantry platform adopts an assembled frame structure, walking mechanisms are respectively arranged on both sides of the bottom of the mobile gantry platform, the rails are laid on the bottom of the mobile gantry platform, and the walking mechanism adopts a synchronous control method.

[0051] In a specific embodiment of the present application, the crusher is driven by a hydraulic motor for vibration excitation, wherein the crusher adjusts the rotation speed of the hydraulic motor in real time to adjust the vibration frequency to reach the resonant frequency of the frozen coal to complete the crushing of the frozen coal.

[0052] In this way, the position, height, stacking shape and other information of the frozen coal in the carriage can be automatically identified, and the crusher can be automatically guided to vibrate and crush the frozen coal. The system controls the feed and movement of the crusher, and makes real-time corrections to the displacement to ensure complete crushing and no damage to the carriage. After the cleaning and recovery system completes the crushing of the frozen coal, the material in the carriage is recovered and the vehicle is cleaned.

[0053] It should be understood that the beneficial effects of this application include:

[0054] 1. The present application is composed of a mobile gantry platform, a lifting device, a frozen coal crushing device, a coal recovery and cleaning device, etc. The lifting device drives the frozen coal crushing device and the coal recovery and cleaning device respectively to achieve the crushing, recovery and cleaning of frozen coal and train compartments. Therefore, by using the present application, manual defreezing can be replaced, thereby reducing the labor intensity of operators and increasing the unloading speed of coal.

[0055] 2. The crushing device can automatically identify the position, height, stacking shape and other information of the frozen coal in the carriage, and automatically guide the crusher to vibrate and crush the frozen coal. The system controls the feed and movement of the crusher and makes real-time corrections to the displacement to ensure complete crushing and no damage to the carriage.

[0056] However, considering that in the actual application of the crusher, if the initial vibration frequency is not preset, it is a long trial and error process to control the vibration frequency by adjusting the speed of the hydraulic motor to reach the resonant frequency of the frozen coal. This will greatly reduce the efficiency of unloading and cleaning. Therefore, in the actual application process, the initial vibration frequency is usually preset, and setting the initial preset value of the vibration frequency depends on the experience and subjectivity of the operator. Due to the inconsistency of the operator's subjectivity and the different experience levels of the operators, the preset results of the initial vibration frequency will be inconsistent, which will increase the debugging time of the vibration frequency.

[0057] In response to the above technical problems, the present application also provides another embodiment. In this embodiment, the technical concept of the present application is: to introduce artificial intelligence technology and image processing technology into the control system to complete the deployment of the initial vibration frequency controller. Specifically, in the initial vibration frequency controller, the frozen coal image collected by the visual guidance system is combined with an intelligent algorithm to perform multi-scale feature mining and fusion on it, so as to learn and capture the multi-scale feature representation of the coal state of the frozen coal image, so as to intelligently decode and restore the recommended value of the initial vibration frequency based on this multi-scale frozen coal state feature representation.

[0058] Figure 2 FIG. 1 is a block diagram of the initial vibration frequency controller in the automatic unloading and cleaning system for frozen coal in a train compartment according to an embodiment of the present application. Figure 2 As shown, the control system 43 also includes an initial vibration frequency controller 431; wherein the initial vibration frequency controller 431 includes: a frozen coal image acquisition module 4311, used to acquire the frozen coal image collected by the visual guidance system; a multi-scale feature extraction module 4312, used to perform multi-scale feature extraction on the frozen coal image to obtain a multi-scale feature map of the coal state; a local feature salient module 4313, used to pass the multi-scale feature map of the coal state through a local feature salient device based on a convolution kernel attention mechanism to obtain a significant multi-scale feature map of the coal state; and, an initial vibration frequency recommended value determination module 4314, used to determine the recommended value of the initial vibration frequency based on the significant multi-scale feature map of the coal state.

[0059] Based on this, in the technical solution of the present application, the specific encoding process of the initial vibration frequency controller includes: first, obtaining the frozen coal image collected by the visual guidance system. Among them, the frozen coal image collected by the visual guidance system can monitor and reflect the state of the coal inside the train compartment, including the freezing condition and the stacking condition. It should be understood that the properties and structures of different coals may be different, such as density, humidity, particle size, etc., and these factors will affect the vibration resonance frequency of coal. At the same time, the frozen coal state may be affected by freezing, and different degrees of freezing will change the stiffness and vibration characteristics of coal, thereby affecting the vibration frequency of coal. In addition, the different stacking conditions of coal in the train compartment may lead to different effects of vibration transmission. For example, loosely stacked coal and tightly stacked coal will have different vibration responses, so the stacking condition of coal will also affect the vibration resonance frequency. Therefore, the frozen coal state information contained in the frozen coal image can be used to judge the state of coal and further determine the initial vibration frequency, so that the initial vibration frequency is closer to the resonance frequency and the debugging time of the vibration frequency is reduced.

[0060] In a specific embodiment of the present application, the multi-scale feature extraction module includes: shallow and deep feature extraction units, which are used to perform shallow feature extraction and deep feature extraction on the frozen coal image respectively to obtain a shallow feature map of the frozen coal state and a deep feature map of the frozen coal state; and a feature fusion attention guidance unit, which is used to pass the shallow feature map of the frozen coal state and the deep feature map of the frozen coal state through a feature fusion attention guidance network to obtain the multi-scale feature map of the coal state.

[0061] Then, the frozen coal image is passed through a frozen coal shallow feature extractor based on a first convolutional neural network model to obtain a frozen coal state shallow feature map; and the frozen coal state shallow feature map is passed through a frozen coal deep feature extractor based on a second convolutional neural network model to obtain a frozen coal state deep feature map. Among them, a convolutional neural network (CNN) is a deep learning neural network that is specifically used to process data with a similar grid structure, such as images. CNN usually includes components such as convolutional layers, pooling layers, and activation layers, and achieves tasks such as image classification and target detection by learning features in images. Here, using a convolutional neural network model to extract shallow feature information and deep feature information of the frozen coal image respectively can fully express the multi-scale state characteristics of the frozen coal image by using shallow features and deep features respectively. Specifically, the shallow features of the frozen coal image, such as edges, textures, and other information, can represent the stacking form of frozen coal and the coal distribution area; the deep features of the frozen coal image can represent more abstract image semantic information in the frozen coal image, such as the physical state information of frozen coal.

[0062] In a specific embodiment of the present application, the shallow and deep feature extraction units are used to: pass the frozen coal image through a frozen coal shallow feature extractor based on a first convolutional neural network model to obtain a shallow feature map of the frozen coal state; and pass the frozen coal state shallow feature map through a frozen coal deep feature extractor based on a second convolutional neural network model to obtain a deep feature map of the frozen coal state.

[0063] Next, the shallow feature map of the frozen coal state and the deep feature map of the frozen coal state are passed through a feature fusion attention guidance network to obtain a multi-scale feature map of the coal state. That is, the feature fusion attention guidance network is used to fuse feature information with different focuses and depths, and an attention mechanism is introduced to enhance attention to important features. Specifically, the specific implementation process of the feature fusion attention guidance network is to perform global average pooling on the channel dimension of the deep feature map of the frozen coal state as a high-level feature. At this time, the high-dimensional feature map will be compressed into a vector representation, and this vector representation has the global receptive field of the deep feature map of the frozen coal state. This vector representation is then used as attention information to fuse with the shallow feature map of the frozen coal state to guide the shallow information in the shallow feature map of the frozen coal state to restore semantic category information, thereby obtaining the multi-scale feature map of the coal state. In this way, the multi-scale feature map of the coal state will have a richer feature representation.

[0064] In a specific embodiment of the present application, the feature fusion attention guiding unit is used to: perform global average pooling on the deep feature map of the frozen coal state along the channel dimension to obtain an attention feature vector; pass the attention feature vector through a fully connected layer to obtain an attention encoding feature vector; perform weighted multiplication on the shallow feature map of the frozen coal state using each eigenvalue in the attention encoding feature vector as a weight to obtain an attention adjustment feature map; and perform positional addition processing on the attention adjustment feature map and the deep feature map of the frozen coal state to obtain the multi-scale feature map of the coal state.

[0065] Subsequently, the multi-scale feature map of the coal state is passed through a local feature salient device based on a convolution kernel attention mechanism to obtain a salient multi-scale feature map of the coal state. Among them, the local feature salient device based on the convolution kernel attention mechanism performs a nonlinear convolution operation on the multi-scale feature map of the coal state through a multi-scale convolution kernel group to extract the associated features in different local spatial neighborhoods in the multi-scale feature map of the coal state, thereby improving the expressiveness and discrimination of the features. Afterwards, the importance of both the channel and the convolution kernel is considered at the same time, and the network is guided by weight information to assign different attentions to convolution kernels of different sizes. In this way, the feature distribution captured by the convolution kernel of useful features is extracted, and the target information of the convolution kernel of useless background features or noise is ignored. Further, the salient multi-scale feature map of the coal state is passed through an initial vibration frequency recommender based on a decoder to obtain a decoded value, and the decoded value is used to represent the recommended value of the initial vibration frequency.

[0066] In a specific embodiment of the present application, the local feature significant module includes: a nonlinear convolution operation unit, which is used to perform a nonlinear convolution operation on the multi-scale feature map of the coal state through a multi-scale convolution kernel group to obtain a first nonlinear convolution feature map, a second nonlinear convolution feature map, a third nonlinear convolution feature map and a fourth nonlinear convolution feature map; a global representation processing unit, which is used to process the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map and the fourth nonlinear convolution feature map with the following global representation formula to obtain a global characterization feature vector; wherein the global representation formula is:

[0067] ;

[0068] ;

[0069] in, , , and are respectively the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map and the fourth nonlinear convolution feature map, is a multi-scale nonlinear convolution fusion feature map, Indicates cascade processing, represents the first eigenvalues, Represents the multi-scale nonlinear convolution fusion feature map The feature matrix is ​​processed by global mean pooling. and They are respectively the first The height and width of the feature matrix, is the first In the feature matrix The invention relates to a method for processing the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map, the fourth nonlinear convolution feature map and the concentrated global characterization feature vector according to the following attention weight application formula to obtain the significant coal state multi-scale feature map; wherein the attention weight application formula is:

[0070] ;

[0071] ;

[0072] ;

[0073] ;

[0074] ;

[0075] ;

[0076] in, , , , Respectively represent the first nonlinear convolution feature map , the second nonlinear convolution feature map , the third nonlinear convolution feature map And the fourth nonlinear convolution feature map The weighted probability vector, represents the condensed global representation feature vector corresponding to the first nonlinear convolution feature map The vector elements of the channel, , and Then, the condensed global representation feature vectors corresponding to the second nonlinear convolution feature map, the third nonlinear convolution feature map, and the fourth nonlinear convolution feature map are represented respectively. The vector elements of the channel, is the significant coal state multi-scale characteristic map, Indicates cascade processing.

[0077] Furthermore, the recommended value determination module of the initial vibration frequency is used to: pass the significant coal state multi-scale feature map through a decoder-based initial vibration frequency recommender to obtain a decoded value, and the decoded value is used to represent the recommended value of the initial vibration frequency.

[0078] In the technical solution of the present application, the feature matrices of the shallow feature map of the frozen coal state and the deep feature map of the frozen coal state respectively express the image semantic features of the frozen coal image based on different convolutional coding depths, and the channel distribution of the shallow feature extractor of the frozen coal state based on the first convolutional neural network model and the deep feature extractor of the frozen coal state based on the second convolutional neural network model is followed between the feature matrices. Therefore, after the shallow feature map of the frozen coal state and the deep feature map of the frozen coal state are fused through the feature fusion attention guidance network, considering the inconsistency of the image semantic space feature distribution corresponding to the shallow feature map of the frozen coal state and the deep feature map of the frozen coal state, the coal state multi-scale feature map will also have the overall spatial feature distribution difference of the feature matrix between the feature matrices under the channel dimension. This overall spatial feature distribution difference of the feature matrix will be further amplified under the action of the local feature salient device based on the convolution kernel attention mechanism, thereby causing local overflow features, affecting the accuracy of the decoding result obtained by the decoder-based initial vibration frequency recommender of the salient coal state multi-scale feature map.

[0079] Accordingly, in a preferred embodiment of the present application, the significant coal state multi-scale feature map is passed through a decoder-based initial vibration frequency recommender to obtain a decoding value, including the following steps: expanding the significant coal state multi-scale feature map into a significant coal state multi-scale feature vector; calculating the self-mean representation matrix and the self-variance representation matrix of the significant coal state multi-scale feature vector, wherein the value of each position of the self-mean representation matrix is ​​the mean of a pair of eigenvalues ​​of two positions of the significant coal state multi-scale feature vector corresponding to the position coordinates, and the value of each position of the self-variance representation matrix is ​​the variance of a pair of eigenvalues ​​of two positions of the significant coal state multi-scale feature vector corresponding to the position coordinates; matrix multiplying the transposed vector of the significant coal state multi-scale feature vector by the self-mean representation matrix to obtain a first significant coal state multi-scale feature intermediate vector and multiplying the self-variance representation matrix by the transposed vector of the significant coal state multi-scale feature vector The method comprises the following steps of: performing matrix multiplication on the first and second significant coal state multi-scale feature intermediate vectors to obtain a second significant coal state multi-scale feature intermediate vector, wherein the significant coal state multi-scale feature vector is a column vector; calculating the point sum of the transposed vectors of the first significant coal state multi-scale feature intermediate vector and the second significant coal state multi-scale feature intermediate vector to obtain a third significant coal state multi-scale feature intermediate vector; calculating the matrix product of the self-mean representation matrix and the self-variance representation matrix, and performing matrix multiplication on the transposed vector of the significant coal state multi-scale feature vector and the matrix product to obtain a fourth significant coal state multi-scale feature intermediate vector; calculating the point sum of the transposed vectors of the third significant coal state multi-scale feature intermediate vector and the fourth significant coal state multi-scale feature intermediate vector to obtain a corrected significant coal state multi-scale feature vector; and passing the corrected significant coal state multi-scale feature vector through the decoder-based initial vibration frequency recommender to obtain the decoded value.

[0080] Based on this, the multimodal load semantic reasoning and representation matrix and the multimodal load semantic reasoning difference representation matrix of the group aggregation statistical evaluation of the local distribution of the characteristic value granularity of the significant coal state multiscale feature vector obtained by expanding the significant coal state multiscale feature map are used as the retrieval and response distribution enhancement of the significant coal state multiscale feature vector, and a reference-free distribution retrieval response framework based on group aggregation and sum-difference feature distribution open domain of the significant coal state multiscale feature vector is constructed to avoid the distribution response redundancy caused by the local overflow characteristics of the significant coal state multiscale feature vector through response superposition, so as to realize the fidelity constraint of the self-aggregation statistical correlation of the response of the significant coal state multiscale feature vector to the target decoding regression domain, and improve the accuracy of the decoding value of the significant coal state multiscale feature vector obtained by the decoder-based initial vibration frequency recommender.

[0081] In summary, the automatic unloading and cleaning system 100 for frozen coal in a train compartment based on the embodiment of the present application is explained, which introduces artificial intelligence technology and image processing technology in the control system to complete the deployment of the initial vibration frequency controller. Specifically, in the initial vibration frequency controller, the frozen coal image collected by the visual guidance system is combined with an intelligent algorithm to perform multi-scale feature mining and fusion on it, so as to learn and capture the multi-scale feature representation of the frozen coal image about the coal state, so as to intelligently decode and restore the recommended value of the initial vibration frequency based on this multi-scale frozen coal state feature representation.

[0082] As described above, the automatic unloading and cleaning system 100 for frozen coal in a train compartment according to an embodiment of the present application can be implemented in various terminal devices, such as a server for automatic unloading and cleaning of frozen coal in a train compartment. In one example, the automatic unloading and cleaning system 100 for frozen coal in a train compartment according to an embodiment of the present application can be integrated into a terminal device as a software module and / or a hardware module. For example, the automatic unloading and cleaning system 100 for frozen coal in a train compartment can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the automatic unloading and cleaning system 100 for frozen coal in a train compartment can also be one of the many hardware modules of the terminal device.

[0083] Alternatively, in another example, the train compartment frozen coal automatic unloading and cleaning system 100 and the terminal device may also be separate devices, and the train compartment frozen coal automatic unloading and cleaning system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0084] In one embodiment of the present application, Figure 3 Flow chart of the method for automatically unloading and cleaning frozen coal from a train compartment according to an embodiment of the present application. Figure 3 As shown, the method for automatically unloading and cleaning frozen coal in a train carriage comprises: 210, setting rails along the travel direction of the railway vehicle; 220, movably installing a mobile gantry platform on the rails; 230, installing a lifting mechanism on the mobile gantry platform, wherein the lifting mechanism comprises a lifting mechanism of a frozen coal crushing system and a lifting mechanism of a cleaning and recovery system; 240, installing a frozen coal crushing system on the lifting mechanism of the frozen coal crushing system, wherein the frozen coal crushing system comprises a crusher and a visual guidance system, a control system is set on the mobile gantry platform, and the visual guidance system is used to identify the status information of the frozen coal in the vehicle. information, using the control system to control the crusher; wherein the control system also includes an initial vibration frequency controller, 250, using the initial vibration frequency controller to obtain a frozen coal image collected by the visual guidance system; 260, performing multi-scale feature extraction on the frozen coal image to obtain a multi-scale feature map of coal state; 270, passing the multi-scale feature map of coal state through a local feature salient device based on a convolution kernel attention mechanism to obtain a significant multi-scale feature map of coal state; and, 280, determining a recommended value of the initial vibration frequency based on the significant multi-scale feature map of coal state.

[0085] Those skilled in the art will appreciate that the specific operations of each step in the above-mentioned method for automatically unloading and cleaning frozen coal from a train compartment have been described in detail above. Figure 1 to Figure 2 The description of the automatic unloading and cleaning system for frozen coal in a train compartment has been described in detail, and therefore, its repeated description will be omitted.

[0086] Figure 4 This is an application scenario diagram of the automatic unloading and cleaning system for frozen coal in a train compartment according to an embodiment of the present application. Figure 4 As shown, in this application scenario, first, a frozen coal image collected by the visual guidance system is obtained (for example, Figure 4 Then, the obtained frozen coal image is input to a server (for example, Figure 4 In S) as shown in , the server can process the frozen coal image based on the automatic unloading and cleaning algorithm of the train compartment frozen coal to determine the recommended value of the initial vibration frequency.

[0087] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0088] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0089] Finally, 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 terminal 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 terminal 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 terminal device including the elements.

[0090] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. An automatic unloading and cleaning system for frozen coal in train compartments. It is characterized in that include: Steel rails set along the direction of travel of railway vehicles; a mobile gantry platform movably mounted on the rails; A lifting mechanism installed on the mobile gantry platform, wherein the lifting mechanism includes a frozen coal crushing system lifting mechanism and a cleaning and recovery system lifting mechanism; and A frozen coal crushing system installed on the lifting mechanism of the frozen coal crushing system, wherein the frozen coal crushing system includes a crusher and a visual guidance system, a control system is arranged on the mobile gantry platform, the visual guidance system is used to identify the state information of the frozen coal in the vehicle, and the control system is used to control the crusher; The mobile gantry platform adopts an assembled frame structure, and walking mechanisms are respectively arranged on both sides of the bottom of the mobile gantry platform. The steel rails are laid on the bottom of the mobile gantry platform, wherein the walking mechanism adopts a synchronous control mode; The crusher is driven by a hydraulic motor to excite the vibration, wherein the crusher adjusts the rotation speed of the hydraulic motor in real time to adjust the vibration frequency to reach the resonance frequency of the frozen coal to complete the crushing of the frozen coal; The control system further includes an initial vibration frequency controller; Wherein, the initial vibration frequency controller comprises: A frozen coal image acquisition module, used to acquire the frozen coal image collected by the visual guidance system; A multi-scale feature extraction module, used for performing multi-scale feature extraction on the frozen coal image to obtain a multi-scale feature map of the coal state; A local feature salient module, used for passing the multi-scale feature map of coal state through a local feature salient device based on a convolution kernel attention mechanism to obtain a salient multi-scale feature map of coal state; and The recommended value determination module of the initial vibration frequency is used to determine the recommended value of the initial vibration frequency based on the significant coal state multi-scale characteristic map.

2. The automatic unloading and cleaning system for frozen coal in a train compartment according to claim 1, It is characterized in that The multi-scale feature extraction module comprises: A shallow and deep feature extraction unit, used to perform shallow feature extraction and deep feature extraction on the frozen coal image to obtain a shallow feature map of the frozen coal state and a deep feature map of the frozen coal state; and The feature fusion attention guiding unit is used to obtain the multi-scale feature map of the coal state by passing the shallow feature map of the frozen coal state and the deep feature map of the frozen coal state through a feature fusion attention guiding network.

3. The automatic unloading and cleaning system for frozen coal in a train compartment according to claim 2, It is characterized in that The shallow and deep feature extraction units are used to: Passing the frozen coal image through a frozen coal shallow feature extractor based on a first convolutional neural network model to obtain a shallow feature map of the frozen coal state; and The shallow feature map of the frozen coal state is passed through a frozen coal deep feature extractor based on a second convolutional neural network model to obtain the deep feature map of the frozen coal state.

4. The automatic unloading and cleaning system for frozen coal in a train compartment according to claim 3, It is characterized in that The feature fusion attention guiding unit is used to: Performing global average pooling along the channel dimension on the deep feature map of the frozen coal state to obtain an attention feature vector; Passing the attention feature vector through a fully connected layer to obtain an attention encoding feature vector; Using each eigenvalue in the attention encoding feature vector as a weight, weighted multiplication is performed on the shallow feature map of the frozen coal state to obtain an attention adjustment feature map; as well as The attention adjustment feature map and the frozen coal state deep feature map are added according to position to obtain the coal state multi-scale feature map.

5. The automatic unloading and cleaning system for frozen coal in a train compartment according to claim 4, It is characterized in that The local feature salient module includes: A nonlinear convolution operation unit, used for performing a nonlinear convolution operation on the multi-scale characteristic map of the coal state through a multi-scale convolution kernel group to obtain a first nonlinear convolution characteristic map, a second nonlinear convolution characteristic map, a third nonlinear convolution characteristic map and a fourth nonlinear convolution characteristic map; A global representation processing unit is used to process the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map and the fourth nonlinear convolution feature map according to the following global representation formula to obtain a global characterization feature vector; wherein the global representation formula is: Among them, U 1 , U 2 , U 3 and U 4 are respectively the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map and the fourth nonlinear convolution feature map, U is a multi-scale nonlinear convolution fusion feature map, represents cascade processing, S c represents the cth eigenvalue in the global characterization eigenvector, F gp (U c ) represents the global mean pooling processing of the c-th feature matrix in the multi-scale nonlinear convolution fusion feature map, H and W are the height and width of the c-th feature matrix in the multi-scale nonlinear convolution fusion feature map, respectively, and U c (i, j) is the eigenvalue at position (i, j) in the c-th feature matrix in the multi-scale nonlinear convolution fusion feature map; An activation processing unit, used for performing activation processing on the global representation feature vector to obtain a concentrated global representation feature vector; An attention weight applying unit is used to process the first nonlinear convolution feature map, the second nonlinear convolution feature map, the third nonlinear convolution feature map, the fourth nonlinear convolution feature map and the concentrated global characterization feature vector using the following attention weight applying formula to obtain the significant coal state multi-scale feature map; wherein the attention weight applying formula is: a c +b c +c c +d c =1 Among them, a c 、b c 、c c ,d c Respectively represent the first nonlinear convolution feature map U 1 , the second nonlinear convolution feature map U 2 , the third nonlinear convolution feature map U 3 And the fourth nonlinear convolution feature map U 4 Acz represents the vector element of the c channel corresponding to the first nonlinear convolution feature map in the concentrated global representation feature vector, Bcz, Ccz and Dcz represent the vector elements of the c channel corresponding to the second nonlinear convolution feature map, the third nonlinear convolution feature map and the fourth nonlinear convolution feature map in the concentrated global representation feature vector, respectively, F c is the significant coal state multi-scale characteristic map, Indicates cascade processing.

6. The automatic unloading and cleaning system for frozen coal in a train compartment according to claim 5, It is characterized in that The recommended value determination module of the initial vibration frequency is used for: The significant coal state multi-scale feature map is passed through a decoder-based initial vibration frequency recommender to obtain a decoded value, and the decoded value is used to represent a recommended value of the initial vibration frequency.

7. A method for automatically unloading and cleaning frozen coal from a train compartment used in the system for automatically unloading and cleaning frozen coal from a train compartment as claimed in any one of claims 1 to 6, It is characterized in that include: The rails are arranged along the direction of travel of the railway vehicles; A movable gantry platform is movably installed on the steel rail; A lifting mechanism is installed on the mobile gantry platform, wherein the lifting mechanism includes a frozen coal crushing system lifting mechanism and a cleaning and recovery system lifting mechanism; The frozen coal crushing system is installed on the lifting mechanism of the frozen coal crushing system, wherein the frozen coal crushing system includes a crusher and a visual guidance system, a control system is arranged on the mobile gantry platform, the visual guidance system is used to identify the state information of the frozen coal in the vehicle, and the control system is used to control the crusher; Among them, the control system also includes an initial vibration frequency controller, which is used to obtain a frozen coal image collected by the visual guidance system; multi-scale feature extraction is performed on the frozen coal image to obtain a multi-scale feature map of coal state; the multi-scale feature map of coal state is passed through a local feature salient device based on a convolution kernel attention mechanism to obtain a significant multi-scale feature map of coal state; and a recommended value of the initial vibration frequency is determined based on the significant multi-scale feature map of coal state.