Intelligent sorting control method for power plant slag removal system based on slag multi-modal identification

Abstract

This invention discloses an intelligent sorting and control method for power plant ash removal systems based on multimodal ash identification. This invention relates to the field of solid waste treatment technology in power plants, solving the problem of distinguishing material morphology solely through simple edge detection or manually set empirical thresholds. This invention uses a dual quantitative judgment rule of the ratio of orthogonal inner diameter lines and the ratio of inner and outer tangent circle areas to transform the contour differences between bottom ash and fly ash into calculable and comparable numerical indicators, eliminating the need for manual experience or complex models. Bottom ash is irregular, with prominent edges and discrete shapes, resulting in significantly larger ratios of inner diameter and inner and outer tangent circle areas. Fly ash particles are small, nearly spherical, and regularly shaped, corresponding to smaller ratios. The two have high distinguishability and clear criteria, enabling millisecond-level rapid classification, effectively reducing the risk of misclassification and mixing, and improving identification consistency and repeatability.

Description

Technical Field

[0001] This invention relates to the field of power plant solid waste treatment technology, specifically to an intelligent sorting and control method for power plant ash removal systems based on multimodal ash and slag identification. Background Technology

[0002] As the core carrier of thermal power generation, power plants generate a large amount of bottom ash and fly ash during their operation. Although both bottom ash and fly ash belong to industrial solid waste, their composition, particle size, and physical morphology differ significantly: bottom ash is mostly blocky, irregular polyhedral particles remaining after combustion, with a large particle size range, sharp edges, and strong internal solidity; fly ash consists of fine particles carried by flue gas, which are nearly spherical or near-spherical, with small particle size, smooth surface, and loose appearance.

[0003] Currently, power plant ash removal systems mostly employ traditional manual sorting or single mechanical screening methods, which present numerous technical bottlenecks. On the one hand, manual sorting relies on experience and judgment, making it susceptible to visual fatigue and environmental dust interference, resulting in low sorting accuracy. It also suffers from high workload, high safety risks, and low efficiency, making it difficult to meet the needs of large-scale, continuous power plant production. On the other hand, single mechanical screening can only achieve coarse particle size classification and cannot accurately identify the material's form and composition. This easily leads to mixing of bottom ash and fly ash, reducing the purity and utilization value of both as building materials and soil conditioners, as well as causing resource waste and increasing environmental treatment costs.

[0004] With the upgrading of industrial intelligence and increasingly stringent environmental protection policies, the industry has placed higher demands on the precision, automation, and digitalization of power plant ash and slag sorting.

[0005] Existing technologies attempt to introduce machine vision for material recognition, but many suffer from defects such as weak anti-interference ability of contour extraction and fuzzy judgment logic. For example, distinguishing material shapes by simply using edge detection or manually setting experience thresholds is easily affected by environmental factors such as material overlap and changes in lighting, resulting in incomplete contour extraction and large fluctuations in judgment results. At the same time, some systems lack component-level verification, and relying solely on morphological features cannot completely distinguish mixed materials with similar shapes, making it difficult to guarantee sorting purity.

[0006] In addition, traditional sorting systems lack effective linkage with warehousing management, and there is a lack of precise directional matching mechanism after material sorting. This can easily lead to the mixing of ash and slag of different qualities in the same silo, making it impossible to achieve the collection of materials of the same quality and the balanced use of silo capacity, which further restricts the overall benefits of solid waste resource utilization in power plants.

[0007] Therefore, there is an urgent need for a ash and slag sorting control method that can achieve accurate material contour extraction, dual verification of morphology and composition, automated sorting execution, and intelligent warehousing matching, in order to solve the problems of low sorting accuracy, insufficient automation, and untraceable data in existing technologies, and promote the upgrading of power plant ash removal systems towards high efficiency, greenness, and intelligence. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides an intelligent sorting and control method for power plant ash removal systems based on multimodal ash and slag recognition, which solves the problem of distinguishing material morphology solely through simple edge detection or manually setting empirical thresholds.

[0009] To achieve the above objectives, the present invention provides the following technical solution: an intelligent sorting and control method for power plant ash removal systems based on multimodal ash identification, comprising the following steps: Step 1: Use machine vision equipment to acquire images of the materials during the conveying process, and use the Sobel algorithm to confirm the contours of the acquired material images, generating image contours associated with the corresponding material images. The specific method is as follows: Machine vision equipment is used to confirm the material image associated with the corresponding material, and the material image is processed into grayscale to confirm the grayscale image: the RGB values ​​associated with different points in the material image are confirmed, and the grayscale value associated with the corresponding point is confirmed by using: R×0.299+G×0.587+B×0.114=grayscale value. Based on the different grayscale values ​​associated with different points, the grayscale image associated with the corresponding material image is generated. Gradient verification is performed on different grayscale points within a grayscale image. The grayscale values ​​associated with the intermediate grayscale point and its surrounding grayscale points are verified. Based on a preset weighting factor, several verified grayscale values ​​are convolved and summed to verify the vertical and horizontal gradients associated with the intermediate grayscale point. The following methods are then employed: Confirm the overall gradient associated with the intermediate grayscale points; Gray points that satisfy the condition that the comprehensive gradient is greater than or equal to Y1 are marked as gradient points; otherwise, no marking is performed. Y1 is a preset value. The contours associated with several consecutive gradient points are connected, and the gradient contours in a closed state are marked as the image contours of the corresponding material image. Step 2: Perform feature verification on the confirmed image contour. Identify mutually perpendicular inner diameter lines or inner and outer tangent circles within the image contour. Locate the line with the largest difference in the ratio of inner diameter lines or the area ratio of the inner and outer tangent circles. Determine whether the material belongs to waste residue based on the ratio. The specific method is as follows: Place the confirmed image contour in a two-dimensional coordinate system, confirm the two-dimensional coordinates associated with different contour points on the image contour, and perform mean processing on several confirmed sets of two-dimensional coordinates to determine the mean coordinates. Based on the specific position of the mean coordinates, mark the associated points in the image contour and record them as the midpoint of the corresponding image contour. Based on the midpoint of the outline marked within the image contour, construct two sets of mutually perpendicular perpendicular lines, with the perpendicular point coinciding with the midpoint of the outline. The endpoints of both sets of perpendicular lines lie on the image contour. Rotate the two sets of mutually perpendicular perpendicular lines around the confirmed midpoint of the outline, ensuring that the endpoints remain on the image contour throughout the rotation. Record the length L1 of the two sets of perpendicular lines at each different rotation stage. i and L2 i Where i represents different rotation stages, and from the two sets of line lengths L1 i and L2 i Extract the maximum and minimum line lengths, and use the formula: maximum line length ÷ minimum line length = Bz i Confirm the line length ratio Bz associated with the corresponding rotation stage. i The maximum value is selected from the different line length ratios confirmed in different rotation stages, and the rotation stage associated with the maximum value is recorded as the standard stage. The line length ratio associated with the standard stage is recorded as the determined ratio. If the determined ratio is ≥1.5, the corresponding material is marked as waste material. If the determined ratio is <1.5, the corresponding material is marked as fly ash material. The specific method for confirming the inner and outer tangent circles within the image contour is as follows: Place the confirmed image contour in a two-dimensional coordinate system and identify the midpoint of the image contour. Using the midpoint of the confirmed contour as the center, confirm the incircle within the image contour and the circumcircle outside the image contour; Mark the confirmed area of ​​the inscribed circle as M1 and the confirmed area of ​​the circumscribed circle as M2. Then, use the formula M2÷M1=Zm to confirm the area ratio Zm. If Zm≥2, then mark the corresponding material as waste material; otherwise, mark it as fly ash material. Step 3: Start the pneumatic three-way material distribution valve and control the flapper action. If the material is determined to be waste residue, switch the valve plate to the slag conveying route. If the material is determined to be fly ash, switch the valve plate to the ash conveying route. Step 4: Rapidly scan the materials along the conveying route using near-infrared sensors to identify the characteristic spectra associated with the corresponding materials. Then, based on the different spectra associated with different material bins throughout history, confirm the spectral characteristics of the corresponding material bins, pinpoint the most similar spectral features of the corresponding materials, and convey them accordingly. The specific method is as follows: If the current conveying route is a slag conveying route, then the different material bins containing waste slag will be marked as bins to be conveyed; if the current conveying route is an ash conveying route, then the different material bins containing fly ash will be marked as bins to be conveyed. Based on the marked warehouses to be transported, determine the spectral characteristics associated with each warehouse, identify the characteristic spectra of the materials transported by each warehouse in the historical process, and average the identified sets of characteristic spectra to lock the average spectrum. Use the identified average spectrum as the spectral characteristics of the corresponding warehouse. Next, based on the characteristic spectrum scanned by the near-infrared sensor, the characteristic spectrum is sequentially compared with the spectral features of different transport chambers to verify similarity, so that the characteristic spectrum and the spectral feature are placed in the same value map, and the characteristic spectrum is controlled to be shifted back and forth. During the shifting process, the maximum overlap state is recorded, and the overlap ratio between the characteristic spectrum and the spectral feature is confirmed from the recorded maximum overlap state: the overlap segment between the characteristic spectrum and the spectral feature is recorded as the standard segment, and the line length ratio of the standard segment in the characteristic spectrum is recorded. The recorded line length ratio is the overlap ratio. Based on the different overlap ratios determined by different delivery warehouses, the delivery warehouse associated with the largest overlap ratio is locked and recorded as the designated warehouse, and the corresponding material is directly delivered to the designated warehouse.

[0010] This invention provides an intelligent sorting and control method for power plant ash removal systems based on multimodal ash identification. Compared with existing technologies, it has the following advantages: By employing a dual quantitative judgment rule of the ratio of orthogonal inner diameter lines and the ratio of the areas of inner and outer tangent circles, the contour differences between bottom ash and fly ash are transformed into calculable and comparable numerical indicators, eliminating the need to rely on manual experience or complex models. Bottom ash is irregular, with prominent edges and discrete shapes, and its ratio of inner diameter and the ratio of the areas of inner and outer tangent circles are significantly larger. Fly ash particles are small, nearly spherical, and have regular shapes, and their corresponding ratios are smaller. The two have high distinguishability and clear judgment criteria, enabling millisecond-level rapid classification, effectively reducing the risk of misclassification and mixing, and improving the consistency and repeatability of identification. Based on the contour recognition results, the pneumatic three-way material distribution valve is directly driven to operate. The valve plate flips quickly to switch the flow direction, realizing online real-time diversion of bottom slag inlet line and fly ash inlet line. No manual operation or downtime adjustment is required throughout the process. It is highly matched with the speed of the conveyor line and does not affect the operation rhythm of the original slag removal system, significantly improving the sorting efficiency and system continuity. By combining near-infrared spectroscopy with rapid scanning, the real-time characteristic spectrum of the material is compared with the historical average spectrum of each storage silo for similarity. The system automatically selects the target silo with the closest composition for transport, achieving precise collection of materials of the same quality. This ensures the stability of the purity of fly ash and bottom ash, while also balancing the material levels in each silo and avoiding quality mixing. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] First Embodiment Please see Figure 1 This application provides an intelligent sorting and control method for power plant ash removal systems based on multimodal ash identification, including the following steps: Step 1: Use machine vision equipment to acquire images of materials during the conveying process, and use the Sobel algorithm to confirm the contours of the acquired material images, generating the image contours associated with the corresponding material images. Specifically, based on the machine vision equipment and the set Sobel algorithm, the overall image corresponding to the material is effectively confirmed. After the overall image is processed into grayscale, the grayscale image associated with the corresponding material can be obtained. Then, according to the set gradient data and the weight factors associated with different positions, the gradient data associated with each point can be effectively confirmed, thereby confirming whether the corresponding point belongs to the gradient point and confirming the gradient contour, completing the overall confirmation process of the edge contour of the corresponding material image. The specific method for generating the image contour corresponding to the material image is as follows: Machine vision equipment is used to confirm the material image associated with the corresponding material, and the material image is processed into grayscale to confirm the grayscale image: the RGB values ​​associated with different points in the material image are confirmed, and the grayscale value associated with the corresponding point is confirmed by using: R×0.299+G×0.587+B×0.114=grayscale value. Based on the different grayscale values ​​associated with different points, the grayscale image associated with the corresponding material image is generated. Gradient verification is performed on different grayscale points within the grayscale image. The grayscale values ​​associated with the central grayscale point and its surrounding (eight groups) grayscale points are verified. Based on a preset weighting factor, the verified groups of grayscale values ​​are convolved and summed to verify the vertical and horizontal gradients associated with the central grayscale point. The following steps are then applied: Confirm the comprehensive gradient associated with the intermediate gray point. Specifically, the preset weight factor value ranges from [-2, 2]. When confirming the vertical and horizontal gradients, the weight factors associated with gray values ​​at different positions are different. Based on the set calculation method, confirm the vertical and horizontal gradients of the corresponding gray point. Since the Sobel algorithm is commonly used to confirm gradient data in existing technologies, it will not be elaborated on here. Gray points that satisfy the condition that the comprehensive gradient is greater than or equal to Y1 are marked as gradient points; otherwise, no marking is performed. Y1 is a preset value, the specific value of which is determined by the operator based on experience. The contours associated with several consecutive gradient points are connected, and the gradient contours in a closed state are marked as the image contours of the corresponding material image. Specifically, after the gradient points around the corresponding material image are determined, the overall edge contour of the corresponding material can be effectively confirmed, thereby effectively confirming the image contour of the corresponding material image. Step 2: Confirm the features of the confirmed image outline, identify the mutually perpendicular inner diameter lines or inner and outer tangent circles within the image outline, locate the line with the largest difference in the ratio of the inner diameter lines or the area ratio of the inner and outer tangent circles, and determine whether the material belongs to waste residue based on the ratio. The specific method for confirming mutually perpendicular inner diameter lines within the image contour is as follows: Place the confirmed image contour in a two-dimensional coordinate system, confirm the two-dimensional coordinates associated with different contour points on the image contour, and perform mean processing on several confirmed sets of two-dimensional coordinates to determine the mean coordinates. Based on the specific position of the mean coordinates, mark the associated points in the image contour and record them as the midpoint of the corresponding image contour. Based on the midpoint of the outline marked within the image contour, construct two sets of mutually perpendicular perpendicular lines, with the perpendicular point coinciding with the midpoint of the outline. The endpoints of both sets of perpendicular lines lie on the image contour. Rotate the two sets of mutually perpendicular perpendicular lines around the confirmed midpoint of the outline, ensuring that the endpoints remain on the image contour throughout the rotation. Record the length L1 of the two sets of perpendicular lines at each different rotation stage. i and L2 i Where i represents different rotation stages, and from the two sets of line lengths L1 i and L2 i Extract the maximum and minimum line lengths, and use the formula: maximum line length ÷ minimum line length = Bz i Confirm the line length ratio Bz associated with the corresponding rotation stage. i The maximum value is selected from the different line length ratios confirmed in different rotation stages, and the rotation stage associated with the maximum value is recorded as the standard stage. The line length ratio associated with the standard stage is recorded as the determined ratio. If the determined ratio is ≥1.5, the corresponding material is marked as waste material; otherwise, it is marked as fly ash material. Specifically, after the overall edge contour of the corresponding material image is determined, two sets of perpendicular lines are constructed within the corresponding edge contour, and both sets of perpendicular lines are in contact with the contour points on the edge contour, that is, the endpoints are always located on the edge contour. By rotating the two sets of perpendicular lines, the two sets of perpendicular lines can be identified within the corresponding contour. Based on the ratio of the lengths of the two sets of perpendicular lines, it can be effectively confirmed whether the corresponding contour is in a regular state. If it is in a regular state, it is not waste residue. The outer contour of waste residue is generally irregular, while fly ash is always in a regular state. Therefore, based on the proportional relationship between the perpendicular lines or the area characteristics of the inner and outer tangent circles, the overall contour characteristics of the corresponding contour can be effectively confirmed. The specific method for confirming the inner and outer tangent circles within the image contour is as follows: Place the confirmed image contour in a two-dimensional coordinate system and identify the midpoint of the image contour. Using the midpoint of the confirmed contour as the center, confirm the incircle within the image contour and the circumcircle outside the image contour; Mark the confirmed area of ​​the inscribed circle as M1 and the confirmed area of ​​the circumscribed circle as M2. Then, use the formula M2÷M1=Zm to confirm the area ratio Zm. If Zm≥2, then mark the corresponding material as waste material; otherwise, mark it as fly ash material. Based on the features of the inner and outer tangent circles of the corresponding contour, the area features associated between the circles can be effectively identified, thereby completing the identification process of waste residue and fly ash. Step 3: Start the pneumatic three-way material distribution valve. If the material is determined to be waste residue, control the flapper action to switch the valve plate to the slag conveying route. If the material is determined to be fly ash, control the flapper action to switch the valve plate to the ash conveying route. Step 4: Perform rapid scanning of materials in the conveying route using near-infrared sensors to identify the characteristic spectra associated with the corresponding materials. Then, based on the different spectra associated with different material bins in the historical process, identify the spectral characteristics of the corresponding material bins, lock in the most similar spectral characteristics of the corresponding materials, and transport them to the corresponding material bins. The specific method for confirming spectral features is as follows: If the current conveying route is a slag conveying route, then the different material bins containing waste slag will be marked as bins to be conveyed; if the current conveying route is an ash conveying route, then the different material bins containing fly ash will be marked as bins to be conveyed. Based on the marked warehouses to be transported, determine the spectral characteristics associated with each warehouse, identify the characteristic spectra of the materials transported by each warehouse in the historical process, and average the identified sets of characteristic spectra to lock the average spectrum. Use the identified average spectrum as the spectral characteristics of the corresponding warehouse. Next, based on the characteristic spectrum scanned by the near-infrared sensor, the characteristic spectrum is sequentially compared with the spectral features of different transport chambers to verify similarity, so that the characteristic spectrum and the spectral feature are placed in the same value map, and the characteristic spectrum is controlled to be shifted back and forth. During the shifting process, the maximum overlap state is recorded, and the overlap ratio between the characteristic spectrum and the spectral feature is confirmed from the recorded maximum overlap state: the overlap segment between the characteristic spectrum and the spectral feature is recorded as the standard segment, and the line length ratio of the standard segment in the characteristic spectrum is recorded. The recorded line length ratio is the overlap ratio. Based on the different overlap ratios determined by different delivery warehouses, the delivery warehouse associated with the largest overlap ratio is locked and recorded as the designated warehouse, and the corresponding material is directly delivered to the designated warehouse; Specifically, based on the infrared spectrum determined by the near-infrared sensor, the different spectral characteristics associated with different material bins are simultaneously determined. A similarity check is used to identify the group of material bins with the closest spectral characteristics to the corresponding materials, and the corresponding materials are transported to the material bin with the highest similarity. This ensures that the materials placed in the corresponding material bins are in a relatively similar state, thereby improving the overall placement effect of the materials.

[0014] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0015] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for intelligent sorting control of a power plant deslagging system based on multi-modal identification of slag, characterized in that, Includes the following steps: Step 1: Use machine vision equipment to acquire images of materials during the conveying process, and use the Sobel algorithm to confirm the contours of the acquired material images, generating the image contours associated with the corresponding material images. Step 2: Confirm the features of the confirmed image outline, identify the mutually perpendicular inner diameter lines or inner and outer tangent circles within the image outline, locate the line with the largest difference in the ratio of the inner diameter lines or the area ratio of the inner and outer tangent circles, and determine whether the material belongs to waste residue based on the ratio. Step 3: Start the pneumatic three-way material distribution valve and control the flapper action. If the material is determined to be waste residue, switch the valve plate to the slag conveying route. If the material is determined to be fly ash, switch the valve plate to the ash conveying route. Step 4: Perform rapid scanning of materials along the conveying route using near-infrared sensors to identify the characteristic spectra associated with the corresponding materials. Then, based on the different spectra associated with different material bins in the historical process, identify the spectral characteristics of the corresponding material bins, lock in the most similar spectral characteristics of the corresponding materials, and convey them.

2. The intelligent sorting and control method for power plant ash removal system based on multimodal ash identification according to claim 1, characterized in that, In step one, the specific method for generating the image contour corresponding to the material image is as follows: Machine vision equipment is used to identify the material image associated with the corresponding material, and the material image is processed into grayscale to confirm the grayscale image: the RGB values ​​associated with different points in the material image are identified, and the grayscale value associated with the corresponding point is confirmed using: R×0.299+G×0.587+B×0.114=grayscale value. Based on the different grayscale values ​​associated with different points, the grayscale image associated with the corresponding material image is generated.

3. The intelligent sorting and control method for power plant ash removal system based on multimodal ash identification according to claim 2, characterized in that, In step one, the specific method for generating the image contour corresponding to the material image also includes: Gradient verification is performed on different grayscale points within a grayscale image. The grayscale values ​​associated with the intermediate grayscale point and its surrounding grayscale points are verified. Based on a preset weighting factor, several verified grayscale values ​​are convolved and summed to verify the vertical and horizontal gradients associated with the intermediate grayscale point. The following methods are then employed: Confirm the overall gradient associated with the intermediate grayscale points; Gray points that satisfy the condition that the comprehensive gradient is greater than or equal to Y1 are marked as gradient points; otherwise, no marking is performed. Y1 is a preset value. The contours associated with several consecutive gradient points are connected, and the gradient contours in a closed state are marked as the image contours of the corresponding material image.

4. The intelligent sorting and control method for power plant ash removal system based on multimodal ash identification according to claim 1, characterized in that, In step two, the specific method for confirming the mutually perpendicular inner diameter lines within the image contour is as follows: Place the confirmed image contour in a two-dimensional coordinate system, confirm the two-dimensional coordinates associated with different contour points on the image contour, and perform mean processing on several confirmed sets of two-dimensional coordinates to determine the mean coordinates. Based on the specific position of the mean coordinates, mark the associated points in the image contour and record them as the midpoint of the corresponding image contour. Based on the midpoint of the outline marked within the image contour, construct two sets of mutually perpendicular perpendicular lines, with the perpendicular point coinciding with the midpoint of the outline. The endpoints of both sets of perpendicular lines lie on the image contour. Rotate the two sets of mutually perpendicular perpendicular lines around the confirmed midpoint of the outline, ensuring that the endpoints remain on the image contour throughout the rotation. Record the length L1 of the two sets of perpendicular lines at each different rotation stage. i and L2 i Where i represents different rotation stages, and from the two sets of line lengths L1 i and L2 i Extract the maximum and minimum line lengths, and use the formula: maximum line length ÷ minimum line length = Bz i Confirm the line length ratio Bz associated with the corresponding rotation stage. i The maximum value is selected from the different line length ratios confirmed in different rotation stages, and the rotation stage associated with the maximum value is recorded as the standard stage. The line length ratio associated with the standard stage is recorded as the determined ratio. If the determined ratio is ≥1.5, the corresponding material is marked as waste material.

5. The intelligent sorting and control method for power plant ash removal system based on multimodal ash identification according to claim 4, characterized in that, If the ratio is determined to be <1.5, the corresponding material will be marked as fly ash material.

6. The intelligent sorting and control method for power plant ash removal system based on multimodal ash identification according to claim 1, characterized in that, In step two, the specific method for confirming the inner and outer tangent circles existing within the image contour is as follows: Place the confirmed image contour in a two-dimensional coordinate system and identify the midpoint of the image contour. Using the midpoint of the confirmed contour as the center, confirm the incircle within the image contour and the circumcircle outside the image contour; Mark the confirmed area of ​​the inscribed circle as M1 and the confirmed area of ​​the circumscribed circle as M2. Then, use the formula M2÷M1=Zm to confirm the area ratio Zm. If Zm≥2, then mark the corresponding material as waste material; otherwise, mark it as fly ash material.

7. The intelligent sorting and control method for power plant ash removal system based on multimodal ash identification according to claim 1, characterized in that, In step four, the specific method for confirming the spectral characteristics of the material warehouse is as follows: If the current conveying route is a slag conveying route, then the different material bins containing waste slag will be marked as bins to be conveyed; if the current conveying route is an ash conveying route, then the different material bins containing fly ash will be marked as bins to be conveyed. Based on the marked warehouses to be transported, determine the spectral characteristics associated with each warehouse, extract the characteristic spectra of the materials transported by each warehouse in the historical process, and average the confirmed sets of characteristic spectra to lock the average spectrum. Use the confirmed average spectrum as the spectral characteristics of the corresponding warehouse.

8. The intelligent sorting and control method for power plant ash removal system based on multimodal ash identification according to claim 7, characterized in that, In step four, the specific method for identifying the most similar spectral features of the corresponding material is as follows: Next, based on the characteristic spectrum scanned by the near-infrared sensor, the characteristic spectrum is sequentially compared with the spectral features of different transport chambers to verify similarity, so that the characteristic spectrum and the spectral feature are placed in the same value map, and the characteristic spectrum is controlled to be shifted back and forth. During the shifting process, the maximum overlap state is recorded, and the overlap ratio between the characteristic spectrum and the spectral feature is confirmed from the recorded maximum overlap state: the overlap segment between the characteristic spectrum and the spectral feature is recorded as the standard segment, and the line length ratio of the standard segment in the characteristic spectrum is recorded. The recorded line length ratio is the overlap ratio. Based on the different overlap ratios determined by different delivery warehouses, the delivery warehouse associated with the largest overlap ratio is locked and recorded as the designated warehouse, and the corresponding material is directly delivered to the designated warehouse.

Images