Automobile coal sampling and measuring method and system

Abstract

The invention relates to the technical field of coal sampling, and particularly discloses an automobile coal sampling measurement method and system, and the method comprises the steps: obtaining a first three-dimensional structure feature map of the morphology of a to-be-sampled coal pile in a target vehicle; dividing a plurality of initial sampling areas according to the three-dimensional structure feature map, randomly selecting at least one initial sampling area to complete first sampling, and generating a first sampling result; generating a second three-dimensional structure feature map of the coal pile morphology after the change of the first sampling according to the sampling features of the sampling machine for the first sampling; if the first sampling result meets the sampling standard, a plurality of second sampling areas are divided according to the second three-dimensional structure feature map, one second sampling area is selected to complete second sampling, a second sampling result is generated, the second sampling result is used for judging whether automobile coal sampling meets the sampling standard or not, and if the automobile coal sampling meets the sampling standard, sampling detection is passed. According to the invention, the accuracy and efficiency of detection can be improved.

Description

Technical Field

[0001] The invention relates to the technical field of coal sampling, and in particular to a method and system for measuring automobile coal sampling. Background Art

[0002] It is crucial that the quality of coal meets the expected standards, as this directly affects energy efficiency, environmental impact and economic benefits. High-quality coal has a higher calorific value and lower impurity content, which means that more energy can be released when burned while producing less pollutants, such as sulfur oxides and particulate matter, which helps to reduce the negative impact on the environment and meet environmental protection requirements. In addition, coal that meets the standards can optimize the combustion process, reduce equipment wear and maintenance costs, and improve the safety and stability of operation. For coal-fired plants that rely on coal as an energy source, it is also a key factor in ensuring production efficiency and economic benefits. In order to ensure that the quality of coal meets the expected standards, by collecting samples from the trucks transporting coal, the physical and chemical properties of coal (such as calorific value, ash content, sulfur content, etc.) can be analyzed to assess whether its quality meets the requirements for specific uses. In addition, this is also an important part of supervising the fairness of transactions, ensuring that the interests of both buyers and sellers are protected, and helping to prevent the occurrence of fraud. The so-called fraud refers to the fact that if the quality of some of the coal in the coal truck is lower than the standard normal value, the coal truck is considered to have fraudulent behavior.

[0003] In the prior art, whether it is a coal sampling method and sampling vehicle with patent application number CN202110470530.2, or the document "Research on Algorithm of Intelligent Sampling Scheme for Mobile Sampling System of Bulk Static Coal" published in Volume 38, Issue 1 of Coal Quality Technology Journal in January 2023, it is necessary to obtain the coal pile morphology through multiple sensors to complete the sampling, and the current sampling times are generally more than 3 times. Each time sampling is performed, it is necessary to obtain the coal pile morphology again through multiple sensors before sampling, and the sampling efficiency is low. Summary of the invention

[0004] The present invention aims to provide a method and system for automobile coal sampling and measurement, which can improve sampling efficiency, improve sampling accuracy, and ensure safety during sampling.

[0005] In order to achieve the above object, the present invention adopts the following technical scheme:

[0006] The first scheme is a method for sampling and measuring coal in an automobile, comprising: obtaining a first three-dimensional structural characteristic map of the morphology of the coal pile to be sampled in the target vehicle; dividing a plurality of initial sampling areas according to the three-dimensional structural characteristic map, randomly selecting at least one initial sampling area to complete the first sampling, and generating the first sampling result; at the same time, generating a second three-dimensional structural characteristic map of the morphology of the coal pile after the first sampling changes according to the sampling characteristics of the sampling machine of the first sampling; if the first sampling result meets the sampling standard, dividing a plurality of second sampling areas according to the second three-dimensional structural characteristic map, selecting a second sampling area to complete the second sampling, and generating the second sampling result, the second sampling result is used to judge whether the automobile coal sampling meets the sampling standard, and if it meets the sampling standard, the sampling test is passed.

[0007] Beneficial effects: Different from the prior art, the prior art uses various sensors to obtain the coal pile morphology after each sampling to ensure the accuracy of the coal pile height and prevent the secondary sampling from being carried out at the adjacent position of the first sampling, which is easily affected by the first sampling and the height of the nearby coal changes accordingly, ensuring the safety of sampling. However, each sampling requires the acquisition of a three-dimensional map of the coal pile morphology before this, which is inefficient.

[0008] First, this first solution performs the corresponding first sampling when acquiring the first three-dimensional structural feature map. At the beginning of the first sampling, the sampling machine has already pre-set the corresponding sampling machine sampling features. Only when the corresponding sampling machine sampling features are pre-set can the corresponding sampling steps be completed. At the same time as the first sampling, the second three-dimensional structural feature map of the coal pile after the first sampling is generated according to the sampling features of the sampling machine. There is no need to re-acquire the corresponding three-dimensional structural feature map of the coal pile through each sensor before the second sampling. In this way, continuous and dynamic sampling can be achieved during sampling. The second sampling can be followed immediately after the first sampling. When detecting the sampling results, continuous detection can be performed, which greatly improves work efficiency and ensures the accuracy and safety of sampling.

[0009] The reason why the prior art did not think of this solution is that, since the general sampling frequency needs to be at least 3 times according to the relevant regulations and standards, the prior art has installed multiple sensors, and the sampling can be completed by allowing multiple sensors to repeatedly obtain the corresponding three-dimensional structural characteristic map of the coal pile according to the previous steps. This first solution does not need to increase equipment, equipment cost, and maintenance cost, and can optimize the corresponding steps, reduce costs, and efficiently complete sampling. Moreover, even if the sampling standard is not passed during the first sampling, the second three-dimensional structural characteristic map generated therefrom will not consume too much acid power. This is because currently one sampling machine generally samples one target vehicle, so the corresponding computing power is used on the sampling machine, and it does not waste much computing power. Instead, the second sampling is directly performed after the first sampling is passed, which can also reduce the additional computing power required for regenerating the three-dimensional characteristic data map using multiple sensors again.

[0010] Secondly, the first sampling is based on the initial sampling area divided by the three-dimensional structural characteristic map to ensure that the sampling points are evenly distributed and representative; the secondary sampling is optimized according to the changes in the coal pile after the first sampling to further improve the accuracy of sampling.

[0011] At the same time, through three-dimensional modeling and intelligent division of sampling areas, repeated measurements and human intervention in traditional sampling are reduced, significantly improving sampling efficiency.

[0012] Moreover, the sampling strategy can be dynamically adjusted according to the actual changes of the coal pile to adapt to coal piles of different shapes and states, thereby improving versatility and flexibility.

[0013] Finally, the secondary sampling results are used for the final judgment, adding a sampling verification step to further ensure the accuracy and reliability of the sampling results, providing strong support for subsequent coal quality monitoring.

[0014] Preferably, a first collection point is set at the entrance and exit of the plant, and the first collection point is used to obtain the chassis height of the target vehicle when the target vehicle is about to leave the plant; a second collection point is set at the coal sampling position, and the second collection point is used to obtain the cross-section of the coal pile to be sampled, and the distances between the coal pile and the driving surface of the target vehicle and the second collection point respectively; the first three-dimensional structural feature map of the morphology of the coal pile to be sampled is generated based on the data collected by the first collection point and the second collection point.

[0015] Beneficial effects: By setting the first collection point at the entrance and exit of the plant to obtain the vehicle chassis height, and setting the second collection point at the coal sampling position to obtain the coal pile cross-section and related distance information, the three-dimensional morphological data of the coal pile can be obtained comprehensively and accurately. This setting not only provides accurate input data for generating the first three-dimensional structural feature map, but also lays the foundation for the subsequent sampling area division and sampling strategy formulation. At the same time, by setting collection points at different locations, the measurement errors caused by changes in vehicle position or movement of coal piles can be effectively reduced, the overall efficiency and accuracy of the sampling system can be improved, and the reliability and safety of the sampling process can be ensured.

[0016] Preferably, the coal pile morphology includes a conical coal pile, a rectangular coal pile, a trapezoidal coal pile, a circular coal pile and a longitudinal coal pile.

[0017] Beneficial effects: The claim clearly divides the coal pile morphology into five types: conical, rectangular, trapezoidal, circular and longitudinal coal piles. The beneficial effect of this classification is that it can accurately perform three-dimensional modeling and sampling area division for coal piles of different shapes. By identifying and distinguishing different coal pile morphologies, the system can adopt a more optimized sampling strategy to ensure the comprehensiveness and representativeness of the sampling, thereby improving the accuracy of the sampling results. In addition, this classification also provides a clearer basis for subsequent coal pile change prediction and sampling path planning, enhances the adaptability and flexibility of coal pile acquisition, and enables it to better cope with complex and changeable coal pile sampling scenarios.

[0018] Preferably, the sampling characteristics of the sampler include sampling position, sampling depth, sampler downward pressure, sample head cross-sectional area, sampler lifting and extension force and sampling volume.

[0019] Beneficial effects: By clarifying the sampling characteristics of the sampler (including sampling position, sampling depth, sampler downforce, sample head cross-sectional area, sampler lifting and extension force, and sampling volume), comprehensive and precise parameter control is provided for the sampling process. This detailed technical parameter setting can ensure the accuracy and consistency of sampling, thereby improving the reliability and representativeness of the sampling results. At the same time, these parameters also provide an important basis for the subsequent prediction of coal pile morphology changes and the optimization of sampling strategies, further improving the intelligence level and adaptability of the sampling system.

[0020] Preferably, the area within the first distance adjacent to the first sampling position is set as the changing area; the changing area is used to simulate the area where the coal pile has changed after the first sampling, and is used to completely replace the area within the first distance adjacent to the sampling position in the first three-dimensional structural characteristic map to form a second three-dimensional structural characteristic map.

[0021] Beneficial effects: By defining the changed area of ​​the coal pile after the first sampling, and using this area to cover the corresponding part in the first three-dimensional structural feature map to form the second three-dimensional structural feature map, the impact of sampling on the morphology of the coal pile can be accurately simulated. This design not only provides more accurate coal pile status information for subsequent sampling, but also optimizes the sampling strategy, improves the scientificity and accuracy of sampling, and reduces the sampling error caused by changes in the coal pile, improving the sampling efficiency and reliability of the results.

[0022] Preferably, a change area generation model is provided, the sampling characteristics of the sampler are used as the first input data, the three-dimensional coordinate data points of the area within a first distance adjacent to the sampling position of the first sampling are obtained, and the three-dimensional coordinate data points are used as the second input data, and the corresponding three-dimensional coordinate data points of the first three-dimensional structural feature map are used as the third input data, and the first input data, the second input data and the third input data are input into the change area generation model to generate corresponding three-dimensional target data, and the three-dimensional target data is used to aggregate to form a change area.

[0023] Beneficial effects: Through multi-dimensional data input, the model can accurately capture the impact of sampling on the coal pile morphology, providing a more realistic reference for subsequent sampling. Based on the generated three-dimensional target data, the system can divide the sampling area more scientifically, avoid sampling deviations caused by changes in the coal pile, and improve the representativeness and accuracy of the sampling. The model can dynamically adjust the change area according to different sampling characteristics and coal pile status, enhancing its adaptive ability, so that it can still operate efficiently under complex working conditions.

[0024] Preferably, a multimodal Transformer architecture is used to set up a change region generation model; the multimodal Transformer architecture includes a convolutional layer for processing the spatial features of three-dimensional point cloud data, a Transformer encoder for capturing the temporal dependencies of sampling parameters, and a cross-attention module for realizing feature alignment of the first input data, the second input data, and the third input data.

[0025] Beneficial effects: It can capture subtle changes in the coal pile during the sampling process, generate high-precision change areas, and provide accurate basis for subsequent sampling. The parallel computing capability and self-attention mechanism of the Transformer architecture can greatly improve the training and reasoning efficiency of the model while reducing the computational complexity. The cross-attention module can achieve deep fusion of different modal data, so that the model can adapt to a variety of coal pile morphologies and sampling conditions, improving the robustness and applicability of the system.

[0026] Preferably, training data for training the change region generation model is acquired according to different coal pile morphologies.

[0027] Beneficial effect: By acquiring training data based on different coal pile morphologies, it is ensured that the change region generation model can learn the change rules of different types of coal piles (such as conical, rectangular, trapezoidal, circular and longitudinal coal piles) during the sampling process. The beneficial effect of this targeted training data acquisition method is that it can significantly improve the adaptability and generalization ability of the model to different coal pile shapes and sampling scenarios, so that it can more accurately predict coal pile changes in practical applications, thereby improving the intelligence level of the sampling system and the accuracy of the sampling results.

[0028] Preferably, the data collected at the first collection point and the second collection point are converted into a plurality of three-dimensional coordinate data points, and the plurality of three-dimensional coordinate data points are fitted to form a three-dimensional structural feature map.

[0029] Beneficial effects: First, through high-precision three-dimensional coordinate data points, the geometric characteristics of the coal pile can be more realistically reflected, providing an accurate basis for the division of sampling areas and the formulation of sampling strategies. Secondly, based on the three-dimensional structural feature map, the sampling area can be divided more scientifically to ensure the comprehensiveness and representativeness of the sampling, thereby improving the efficiency of sampling and the accuracy of the results. Finally, it can be used for coal piles of different shapes and sizes, which improves the versatility and flexibility of the system and enables it to better adapt to complex sampling environments.

[0030] The second scheme is applied to a method for sampling and measuring coal in an automobile as described in the first scheme, wherein the data acquisition module is used to obtain the cross section of the coal pile of the target vehicle and the actual height of the coal pile as the first acquisition data; the three-dimensional modeling module converts the first acquisition data into a first three-dimensional structural feature map of the coal pile; the first sampling module divides at least one initial sampling area according to the first three-dimensional structural feature map of the coal pile and completes the first sampling; the deformation prediction module sets a change area generation model, takes the sampling characteristics of the sampler as the first input data, obtains the three-dimensional coordinate data points of the area within a first distance adjacent to the sampling position of the first sampling, and takes the three-dimensional coordinate data points as the second input data, and the corresponding three-dimensional coordinate data points of the first three-dimensional structural feature map as the third input data, and generates corresponding three-dimensional target data by inputting the first input data, the second input data and the third input data into the change area generation model, and the three-dimensional target data is used to aggregate to form a change area, and the change area is used to simulate the change of the coal pile after the first sampling; the secondary sampling module forms a second three-dimensional structural feature map with the change of the coal pile after sampling, divides a plurality of second sampling areas according to the second three-dimensional structural feature map, selects a second sampling area to complete the second sampling, and generates a second sampling result; the secondary sampling judgment module is used to judge whether the second sampling result meets the standard.

[0031] Beneficial effects: Similar to the beneficial effects of an automobile coal sampling and measurement method applied to the first scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A schematic diagram of a method for sampling and measuring automobile coal in Example 1;

[0033] Figure 2 This is a schematic diagram of collecting data information of the target vehicle at the second collection point. DETAILED DESCRIPTION

[0034] The following embodiments of the technical solution of the present invention are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and are therefore only used as examples, and cannot be used to limit the protection scope of the present invention.

[0035] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by those skilled in the art to which the present invention belongs.

[0036] Embodiment 1

[0037] like Figure 1 As shown, this embodiment provides a method for sampling and measuring coal in an automobile, which specifically includes:

[0038] In the coal-carrying car coal sampling process of a coal-fired power plant, when the car carrying coal drives to the exit of the plant, the car is in an empty compartment state, and the first collection point set here starts working. The first collection point is equipped with a high-precision laser sensor and image recognition equipment. The laser sensor accurately measures the chassis height of the target vehicle by emitting a laser beam and receiving reflected light. At the same time, the image recognition equipment originally set up at the exit of the plant takes a photo to identify the target vehicle and records the basic information of the vehicle, such as the license plate number and model, for subsequent tracking and data matching. At the first collection point, a first support frame is installed, on which a high-precision laser sensor and image recognition equipment are fixedly installed. The laser sensor faces the target vehicle from top to bottom.

[0039] When the target vehicle drives to the coal sampling location, the second collection point comes into play. A variety of measuring equipment is installed here, including a 3D laser scanner and an ultrasonic rangefinder. The 3D laser scanner emits a laser beam at a high frequency to scan the entire coal pile to be sampled, such as Figure 2As shown, the cross-sectional information of the coal pile is obtained. The cross-sectional information of the coal pile is parallel to the ground. The purple box is the sampling area. The target vehicle is located in the sampling area. The green box is set as the cross-sectional size of the carriage of the target vehicle. The red box is the area where the coal pile is placed. The three-dimensional laser scanner can accurately measure the cross-sectional information of the coal pile and the contour information of each angle of the cross-sectional area. The ultrasonic rangefinder is used to measure the distance between the coal pile, the driving surface of the target vehicle and the second collection point. By emitting ultrasonic waves and receiving reflected waves, it can accurately calculate these distance data, such as the distance from a certain position on the top of the coal pile to the second collection point. The height of the second collection point from the ground is the first fixed height, and then the actual chassis height of the target vehicle collected by the first collection point. In the second collection point, a second support frame is fixedly installed, and a sampler is installed on the second support frame. The sampler is used to sample coal from the target vehicle. The first support frame is also fixedly installed with a first support frame, and an ultrasonic rangefinder, a three-dimensional laser scanner and a monitoring camera sensor are fixedly installed on the first support frame. The ultrasonic rangefinder, the three-dimensional laser scanner and the monitoring camera sensor are all located above the sampler, and can clearly obtain the coal pile morphology of the target vehicle. When the target vehicle enters the factory for the first time, the conventional chassis height corresponding to its corresponding model can be used as the corresponding collection data.

[0040] After obtaining these data, they are transmitted to the data processing center. In the data processing center, the actual height of the coal pile can be obtained by calculating the first fixed height, the actual chassis height of the target vehicle, and the distance from a certain position on the top of the coal pile to the second collection point. And each data point is converted into multiple three-dimensional coordinate data points. Each three-dimensional coordinate data point is assigned a corresponding three-dimensional coordinate, such as (x, y, z). By fitting a large number of three-dimensional coordinate data points, the first three-dimensional structural feature map of the morphology of the coal pile to be sampled is finally formed. This three-dimensional structural feature map shows the shape, size and location information of the coal pile in an intuitive and accurate way, providing an important basic basis for subsequent sampling work.

[0041] According to the generated three-dimensional structural feature map, the type of coal pile morphology is judged, and the coal pile morphology includes conical coal pile, rectangular coal pile, trapezoidal coal pile, circular coal pile and longitudinal coal pile. Through the regional division algorithm, the coal pile is divided into multiple initial sampling areas, which can be divided equally or according to the coal pile morphology. The division of these areas can fully consider the shape, size and coal characteristics of different parts of the coal pile to ensure the comprehensiveness and representativeness of the sampling. For example, for a larger rectangular coal pile, it can be divided into nine initial sampling areas according to its coal pile cross section.

[0042] Subsequently, the data processing center uses a random number generator to randomly select at least one area from these initial sampling areas for the first sampling. Assume that the area located at the top of the upper right corner of the coal pile is randomly selected. At this time, when the sampler starts working, the corresponding sampling characteristics of the sampler have been preset, and the corresponding sampling is performed based on the sampling characteristics of the sampler. The sampling characteristics of the sampler include sampling position, sampling depth, sampler down pressure, sample head cross-sectional area, sampler lifting and extension force, and sampling volume. The basic information of the coal sample collected by the sampler, such as the weight and appearance characteristics of the coal sample, as well as the data obtained from the preliminary analysis of the coal sample, such as the preliminary calorific value and ash content of the coal sample.

[0043] At the same time as the first sampling, according to the sampling characteristics of the sampler, the data processing center generates a second three-dimensional structural characteristic map of the coal pile after the first sampling. Specifically, the sampling position is used as a dot, and the area within the first distance is set as the change area. This first distance can be set according to the actual situation, and the first distance is 0.5 meters. The change area is used to simulate the area where the coal pile changes after the first sampling. According to the sampling depth, lifting force and other parameters of the sampler, the deformation of the coal pile after sampling is calculated, and then the three-dimensional coordinate point data corresponding to the change area is used to cover the area within the first distance adjacent to the sampling position in the first three-dimensional structural characteristic map, thereby forming a second three-dimensional structural characteristic map. This second three-dimensional structural characteristic map can accurately reflect the state of the coal pile after the first sampling, and provide a more accurate reference for subsequent sampling work.

[0044] Specifically, in the construction of the change area generation model, data collection and feature engineering are completed. The input data includes at least first input data and second input data. The first input data includes sampling depth, downward pressure, lifting force, and cross-sectional area of ​​the sample head; the second input data is the three-dimensional coordinate point cloud data within 0.5m around the sampling position (including X / Y / Z axis spatial coordinates and point cloud density), and the corresponding three-dimensional coordinate data points of the first three-dimensional structural feature map are used as the third input data. On this basis, the spatiotemporal feature matrix can be established by integrating the physical property parameters of the coal pile (such as density, moisture content, etc., which are correction coefficients fitted through historical data), which can be the correlation between the sampling time series and the spatial position.

[0045] When establishing the spatiotemporal characteristic matrix of coal pile changes, the three-dimensional target data corresponding to the change area can be expressed as:

[0046] Z t =c+λ 1 Z t-1 +λ 2 Z t-2 +...+λ n Z t-d +a; where Z tis the Z-axis (coal pile height change) vector corresponding to a certain two-dimensional coordinate point in the change area, Zt contains a vector of λ variables and represents the value at time t (equivalent to the sampling time). In this embodiment, the sampling time is the time from the sampler extending to the complete retraction for sampling. 1 ,...,λ n is the regression coefficient matrix, which indicates the weight value of different subsequent steps d. a is usually the error coefficient, which is generally taken as a vector modulus value of 0.2 to 0.5.

[0047] The multimodal Transformer architecture is adopted, including a convolutional layer for processing the spatial features of 3D point cloud data, a Transformer encoder for capturing the temporal dependencies of sampling parameters, and a cross-attention module for feature alignment of three modal data. The 3D deformation error function is used as the loss function, which includes position offset error, volume change error, and surface normal error. Physical constraints can also be introduced, such as regularization terms based on particle flow dynamics equations.

[0048] When preparing training data, training data covering five typical coal pile morphologies, such as conical coal pile, rectangular coal pile, trapezoidal coal pile, circular coal pile and longitudinal coal pile, are constructed, including equally dividing the initial sampling area in each coal pile, collecting the usage of the existing samplers in use according to different depths and different areas, and assuming that the downforce and lifting force generated by the sampler each time it works are the same data to simplify the construction of training data. At the same time, the Monte Carlo method is used to generate virtual sampling scenes to enhance the generalization ability of the model. In the output content, the coordinate offset matrix of the change area generation model, the change matrix of the coal pile surface normal vector field, and the three-dimensional target data corresponding to the change area after the first sampling are output. In the training data collection, the existing technology can be used to generate the three-dimensional data of the three-dimensional structural feature map before the first sampling, and after the first sampling is completed, the three-dimensional data of the three-dimensional structural feature map after the first sampling is generated. By comparing each data point, the difference between each three-dimensional data point on the same X and Y axis on the Z axis is used as a change point, and the change points are aggregated to form a change area. In this embodiment, the cross section of the coal pile is the plane where the X and Y axes are located, and the height of the coal pile is the Z axis. In the training data source and collection method, the three-dimensional structural feature map of the coal pile is collected by a high-precision three-dimensional laser scanner, an ultrasonic rangefinder and a sampler sensor. The coal pile morphology data before and after the first sampling is recorded in detail, including the cross section, height, sampling position, sampling depth, sampler downforce, lifting force and other parameters of the coal pile. These data provide a real sampling scene for the change area generation model to ensure that the model can learn the physical change laws in the actual sampling process. At the same time, computer simulation technology is used to generate a virtual sampling scene based on the known physical properties of the coal pile (such as density, moisture content, particle size distribution, etc.) and the working principle of the sampler. Different combinations of coal pile morphology, sampling positions and sampling parameters are randomly generated by the Monte Carlo method, thereby generating a large number of virtual training samples. These simulated data can enhance the generalization ability of the model, enabling it to adapt to a wider range of sampling scenarios.

[0049] Specifically, it includes the three-dimensional structural characteristic map of the coal pile before the first sampling and the three-dimensional structural characteristic map of the coal pile after the first sampling. Each data point contains three-dimensional coordinates (X, Y, Z) and related physical properties (such as density, moisture content, etc.). These data are used to describe the shape, size and spatial distribution of the coal pile. By comparing the three-dimensional structural characteristic maps before and after the first sampling, the change information of the coal pile morphology is extracted. This change information includes the change in the height of the coal pile and the change in the surface normal vector. These data are used as the output target of the model to train the model to learn the physical change law of the coal pile during the sampling process. For the sampler parameter data, standardization is performed so that it can be compared and analyzed under the same dimension. For each training sample, the change area of ​​the coal pile morphology after the first sampling is marked, including the boundary of the change area, the degree of change (Z-axis height change) and the three-dimensional coordinate data point after the change.

[0050] The beneficial effects of this embodiment

[0051] First, in the prior art, each sampling requires multiple sensors to re-acquire the coal pile morphology, which is inefficient. The present invention directly generates a second three-dimensional structural feature map at the same time as the first sampling, which is directly used for the division of subsequent sampling areas and sampling decisions, avoiding the step of repeatedly measuring the coal pile morphology and significantly improving the sampling efficiency.

[0052] Secondly, after the first sampling, the system can quickly generate a changed coal pile morphology map and divide the second sampling area accordingly, reducing the waiting time and equipment operation time during the sampling process, and further improving the overall efficiency of the sampling process. In addition, the first three-dimensional structural feature map of the coal pile is obtained through high-precision laser sensors and three-dimensional laser scanners, which can accurately reflect the shape, size and spatial distribution information of the coal pile. This high-precision three-dimensional modeling provides a more accurate basis for the division of sampling areas and the formulation of sampling strategies, avoiding sampling deviations caused by inaccurate coal pile morphology.

[0053] At the same time, according to the different morphologies of the coal pile (such as cone, rectangle, trapezoid, etc.), the regional division algorithm is used to divide the coal pile into multiple initial sampling areas, and the second sampling area is re-divided according to the changes after the first sampling. This scientific regional division method can ensure the comprehensiveness and representativeness of the sampling, and avoid inaccurate sampling results caused by improper selection of sampling areas. Then, the change region generation model of the multimodal Transformer architecture can accurately predict the changes of the coal pile after sampling based on the sampling characteristics of the sampler and the physical properties of the coal pile. The generated second three-dimensional structural feature map can truly reflect the state of the coal pile after sampling, providing an accurate reference for subsequent sampling, thereby improving the accuracy of sampling.

[0054] Next, various coal pile morphologies such as conical coal piles, rectangular coal piles, trapezoidal coal piles, annular coal piles and longitudinal coal piles were considered, and corresponding sampling strategies were designed for coal piles of different morphologies. The sampling strategy includes the division of sampling areas and the training data of the variable area generation model. This enables the system to adapt to coal piles of different shapes and sizes and has a wide range of applicability. At the same time, virtual sampling scenes are generated by the Monte Carlo method and used as part of the training data, which enhances the generalization ability of the variable area generation model. The model can learn more diverse sampling scenes and coal pile change laws, so as to better adapt to different sampling conditions and coal pile characteristics in practical applications and improve the robustness of the system.

[0055] In addition, since the steps of repeatedly measuring the coal pile morphology are reduced, the number of times the sampling equipment is used is reduced accordingly, which not only reduces the loss and maintenance cost of the equipment, but also reduces the possibility of equipment failure and improves the reliability of the system. At the same time, through the intelligent decision-making module and the change area generation model, the system can generate the optimal sampling path and strategy according to the real-time status of the coal pile and the sampling target. This avoids unnecessary sampling operations and equipment movement, further reducing the sampling cost.

[0056] Finally, the use of high-precision laser sensors and image recognition equipment can accurately measure the chassis height of the target vehicle and the actual height of the coal pile, avoiding the risk of collision between the sampler and the vehicle or coal pile due to insufficient equipment accuracy, and improving the safety of the sampling process.

[0057] Embodiment 2

[0058] The present embodiment provides a vehicle coal sampling and measurement system, which is applied to a vehicle coal sampling and measurement method described in the first embodiment, including: a data acquisition module, which is used to obtain the coal pile cross section of the target vehicle and the actual height of the coal pile as the first acquisition data; a three-dimensional modeling module, which converts the first acquisition data into a first coal pile three-dimensional structural feature map; a first sampling module, which divides at least one initial sampling area according to the first coal pile three-dimensional structural feature map, and completes the first sampling; a deformation prediction module, which sets a change area generation model, takes the sampling characteristics of the sampler as the first input data, obtains the three-dimensional coordinate data points of the area within a first distance adjacent to the sampling position of the first sampling, and converts the three-dimensional coordinate data points into the first sampling area; The data point is used as the second input data, and the corresponding three-dimensional coordinate data point of the first three-dimensional structural feature map is used as the third input data. The first input data, the second input data and the third input data are input into the change area generation model to generate corresponding three-dimensional target data. The three-dimensional target data is used to aggregate to form a change area, and the change area is used to simulate the change of the coal pile after the first sampling; a secondary sampling module forms a second three-dimensional structural feature map with the change of the coal pile after sampling, divides a plurality of second sampling areas according to the second three-dimensional structural feature map, selects a second sampling area to complete the second sampling, and generates a second sampling result; a secondary sampling judgment module is used to judge whether the second sampling result meets the standard.

[0059] This embodiment is a brief description, and specific reference may be made to Embodiment 1, which shall be taken as a corresponding basis.

[0060] In this specification, a lot of specific details are described. However, it is understood that embodiments of the present invention can be put into practice without these specific details. In some examples, known methods, systems and techniques are not shown in detail, so as not to obscure the understanding of this specification. In the description of this specification, the description of reference terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, methods, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of this specification.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A method for sampling and measuring coal from an automobile, characterized in that: include: Obtain the first three-dimensional structural feature map of the coal pile to be sampled in the target vehicle; Divide a plurality of initial sampling areas according to the three-dimensional structural characteristic map, randomly select at least one initial sampling area to complete the first sampling, and generate the first sampling result; and at the same time, generate a second three-dimensional structural characteristic map of the coal pile morphology after the first sampling changes according to the sampling characteristics of the sampling machine of the first sampling; If the first sampling result meets the sampling standard, multiple second sampling areas are divided according to the second three-dimensional structural characteristic map, one second sampling area is selected to complete the second sampling, and a second sampling result is generated. The second sampling result is used to determine whether the automobile coal sampling meets the sampling standard. If it meets the sampling standard, the sampling test is passed.

2. The method for sampling and measuring coal from an automobile according to claim 1, characterized in that: A first collection point is set at the entrance and exit of the factory area, and the first collection point is used to obtain the chassis height of the target vehicle when the target vehicle is about to leave the factory area; A second collection point is set at the coal sampling position, and the second collection point is used to obtain the cross section of the coal pile to be sampled, and the distances between the coal pile, the target vehicle driving surface and the second collection point respectively; A first three-dimensional structural characteristic map of the coal pile to be sampled is generated based on the data collected at the first collection point and the second collection point.

3. The method for sampling and measuring coal from an automobile according to claim 1, characterized in that: Coal pile shapes include conical coal piles, rectangular coal piles, trapezoidal coal piles, circular coal piles and longitudinal coal piles.

4. The method for sampling and measuring coal from an automobile according to claim 1, characterized in that: The sampling characteristics of the sampler include sampling position, sampling depth, downward pressure of the sampler, cross-sectional area of ​​the sample head, lifting and extension force of the sampler and sampling volume.

5. The method for sampling and measuring coal in an automobile according to claim 1, characterized in that: The area within the first distance adjacent to the first sampling position is set as the changing area; the changing area is used to simulate the area where the coal pile has changed after the first sampling, and is used to completely replace the area within the first distance adjacent to the sampling position in the first three-dimensional structural characteristic map to form a second three-dimensional structural characteristic map.

6. The method for sampling and measuring coal from an automobile according to claim 5, characterized in that: A change region generation model is provided, and the sampling characteristics of the sampler are used as the first input data to obtain the three-dimensional coordinate data points of the area within a first distance adjacent to the sampling position of the first sampling, and the three-dimensional coordinate data points are used as the second input data, and the corresponding three-dimensional coordinate data points of the first three-dimensional structural feature map are used as the third input data. By inputting the first input data, the second input data and the third input data into the change region generation model, corresponding three-dimensional target data are generated, and the three-dimensional target data are used to aggregate to form a change region.

7. The method for sampling and measuring coal from an automobile according to claim 6, characterized in that: A multimodal Transformer architecture is used to set up a change region generation model; the multimodal Transformer architecture includes a convolutional layer for processing spatial features of three-dimensional point cloud data, a Transformer encoder for capturing temporal dependencies of sampling parameters, and a cross-attention module for realizing feature alignment of first input data, second input data, and third input data.

8. The method for sampling and measuring coal from an automobile according to claim 6, characterized in that: The training data for training the change area generation model is obtained according to the different coal pile morphologies.

9. The method for sampling and measuring coal from an automobile according to claim 2, characterized in that: The data collected at the first collection point and the second collection point are converted into a plurality of three-dimensional coordinate data points, and the plurality of three-dimensional coordinate data points are fitted to form a three-dimensional structural feature map.

10. A vehicle coal sampling and measuring system according to claim 1, characterized in that: A method for sampling and measuring coal in an automobile as described in any one of claims 1 to 9, comprising: A data acquisition module, used to obtain the coal pile cross section of the target vehicle and the actual height of the coal pile as the first acquisition data; The 3D modeling module converts the first collected data into the first 3D structural feature map of the coal pile; The first sampling module divides at least one initial sampling area according to the first three-dimensional structural characteristic map of the coal pile and completes the first sampling; The deformation prediction module sets a change region generation model, takes the sampling characteristics of the sampler as the first input data, obtains the three-dimensional coordinate data points of the region within the first distance adjacent to the sampling position of the first sampling, and takes the three-dimensional coordinate data points as the second input data, and the corresponding three-dimensional coordinate data points of the first three-dimensional structural feature map as the third input data, and generates corresponding three-dimensional target data by inputting the first input data, the second input data and the third input data into the change region generation model, and the three-dimensional target data is used to aggregate to form a change region, and the change region is used to simulate the change of the coal pile after the first sampling; The secondary sampling module forms a second three-dimensional structural characteristic map based on the changes of the coal pile after sampling, divides a plurality of second sampling areas according to the second three-dimensional structural characteristic map, selects one second sampling area to complete the second sampling, and generates a second sampling result; The secondary sampling judgment module is used to judge whether the second sampling result meets the standard.

Images