CN115375675A - Coal quality detection method based on image data - Google Patents

Abstract

The invention relates to the technical field of image data processing, in particular to a coal quality detection method based on image data, which comprises the following steps: acquiring a target surface image and a reflection spectrum curve chart of coal to be detected; carrying out self-adaptive threshold edge detection on the target surface image; determining the coal granularity corresponding to the coal to be detected; screening sample coal information meeting the particle condition from the sample coal information set; screening target coal information meeting similar conditions from the target coal information set; determining the sample coal ash content included by the similar coal information as the target coal ash content of the coal to be detected; and generating target coal quality information. According to the invention, the technical problems of low efficiency and accuracy of coal quality detection are solved by image processing of the target surface image and the reflection spectrum curve graph, the efficiency and accuracy of coal quality detection are improved, and the method is mainly applied to coal quality detection.

Description

Coal quality detection method based on image data

Technical Field

The invention relates to the technical field of image data processing, in particular to a coal quality detection method based on image data.

Background

At a certain temperature, coal is completely combusted, and the percentage of the residual solid substances in the coal mass is often called coal ash. The coal ash content is an important measurement index for indicating the content of inorganic matters in coal, the larger the coal ash content is, the more inorganic matters are contained in the coal, and the poorer the quality of the coal is, and on the contrary, the smaller the coal ash content is, the less inorganic matters are contained in the coal, and the better the quality of the coal is. At present, when the quality of coal is detected, the method generally adopted is as follows: and judging the coal quality according to the coal ash content of the coal. The coal ash content of coal is usually detected by a burning method, which is a coal ash content detection method widely applied. In the prior art, a method for detecting ash content of coal based on image processing is also available, which comprises the following steps: firstly, a coal image of coal to be detected is obtained, the coal image is preprocessed to obtain a target image, and then the coal ash content of the coal to be detected is determined according to the average gray value of the target image and a pre-trained coal ash content detection model. The coal ash content detection model can be a model established according to the mapping relation between the average training gray value and the coal ash content corresponding to a plurality of training coals. The training coal may be coal of known coal ash content. The average training grayscale value may be an average grayscale value of a training coal image.

However, when the above-described manner is adopted, there are often technical problems as follows:

firstly, when a coal ash component of coal is detected by a burning method, the implementation process is often complex, the analysis period is often long, the efficiency of detecting the coal quality is often low, the detection result is often influenced by manual operation, and when the manual operation is not proper, the detection result obtained by detecting the coal quality is often inaccurate, so that the accuracy of detecting the coal quality is low;

secondly, when the method for detecting the coal ash based on the image processing is adopted to detect the coal ash, only the mapping relation between the average gray value of the target image and the coal ash corresponding to the coal to be detected is usually considered, however, the determination of the coal ash corresponding to the coal to be detected is usually influenced by various factors and not only by the average gray value, and therefore, when the method is adopted to detect the coal ash, the accuracy of detecting the coal quality is often low.

Disclosure of Invention

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

The invention provides a coal quality detection method based on image data, and aims to solve the technical problem that the efficiency and the accuracy of coal quality detection are low.

The invention provides a coal quality detection method based on image data, which comprises the following steps:

acquiring a target surface image and a reflection spectrum curve chart of coal to be detected;

carrying out self-adaptive threshold edge detection on the target surface image to obtain a target texture feature and a coal particle outline set;

determining the coal granularity corresponding to the coal to be detected according to the target texture characteristics and the coal particle outline set;

according to the coal granularity, screening sample coal information meeting the particle condition from a pre-obtained sample coal information set, and obtaining a target coal information set as target coal information, wherein the sample coal information in the sample coal information set comprises: sample coal ash content, a sample reflection spectrum curve graph and sample coal granularity, wherein the particle condition is that the sample coal granularity included in the sample coal information is equal to the coal granularity;

screening target coal information meeting similar conditions from the target coal information set according to the reflection spectrum curve graph, and a sample reflection spectrum curve graph and sample coal ash content included in each target coal information in the target coal information set, and taking the target coal information as similar coal information;

determining the sample coal ash content included in the similar coal information as the target coal ash content of the coal to be detected;

and generating target coal quality information corresponding to the coal to be detected according to the target coal ash content.

Further, the performing adaptive threshold edge detection on the target surface image to obtain a target texture feature and a coal particle contour set includes:

determining a target gradient histogram corresponding to the target surface image according to the target surface image;

determining a high threshold of adaptive threshold edge detection according to the target gradient histogram;

determining a numerical value which is different from the high threshold value by a preset numerical value as a low threshold value of the self-adaptive threshold value edge detection;

according to the high threshold and the low threshold, performing edge detection on the target surface image to obtain an edge image, wherein the edge image comprises: collecting edges;

determining reference texture features according to the edge images;

determining fine edge features according to the target surface image and the edge set;

determining complete edge features according to the edge set;

determining an edge detection effect index according to the reference texture feature, the fine edge feature and the edge integrity feature;

determining a target low threshold according to the edge detection effect index and a preset effect index threshold;

determining the reference texture feature corresponding to the target low threshold as a target texture feature;

and performing edge detection on the target surface image according to the high threshold and the target low threshold to obtain a coal particle outline set.

Further, the determining a target low threshold according to the edge detection effect index and a preset effect index threshold includes:

when the edge detection effect index is larger than the effect index threshold, determining a low threshold as a target low threshold;

when the low threshold is greater than 0 and the edge detection effect index is less than or equal to the effect index threshold, updating the low threshold to be the low threshold minus a preset value;

according to the high threshold and the updated low threshold, performing edge detection on the target surface image to obtain an updated edge image, wherein updating the edge image comprises: updating the edge set;

determining an updated reference texture feature according to the updated edge image;

determining updated tiny edge features according to the target surface image and the updated edge set;

determining the complete characteristics of the update edge according to the update edge set;

determining an updated edge detection effect index according to the updated reference texture feature, the updated fine edge feature and the updated edge integrity feature;

when the updated low threshold is greater than 0 and the updated edge detection effect index is less than or equal to the effect index threshold, taking the updated low threshold as the low threshold, repeating the steps until the updated edge detection effect index is greater than the effect index threshold or the updated low threshold is not greater than 0, and stopping repeating;

when the updated edge detection effect index is larger than the effect index threshold, determining the updated low threshold as a target low threshold;

and when the updated low threshold value is not more than 0, determining the edge detection effect index and the maximum value of all the updated edge detection effect indexes as a target detection effect index, and determining the low threshold value corresponding to the target detection effect index as a target low threshold value.

Further, the determining the reference texture feature according to the edge image includes:

determining a reference gray level co-occurrence matrix according to the edge image;

determining energy and inverse variance according to the reference gray level co-occurrence matrix;

and determining the reference texture features according to the energy and the inverse variance.

Further, the determining fine edge features according to the target surface image and the edge set includes:

determining a gray level histogram according to the target surface image;

performing threshold segmentation on the gray level histogram to obtain a gray level threshold;

dividing the target surface image according to a gray threshold value to obtain a highlight area set;

screening out a fine edge set from the edge set according to the highlight area set;

and determining the fine edge characteristics according to the highlight region set and the fine edge set.

Further, the determining the edge integrity feature according to the edge set includes:

dividing edges in the edge set to obtain a closed edge set and a non-closed edge set;

and determining the edge integrity feature according to the number of the non-closed edges in the non-closed edge set, the number and the area of the closed edges in the closed edge set and the mean value of the areas of the closed edges in the closed edge set.

Further, the determining the coal granularity corresponding to the coal to be detected according to the target texture feature and the coal particle contour set includes:

determining the area mean value of the coal particle outline according to the coal particle outline set;

and determining the coal granularity according to the coal particle outline area mean value and the target texture feature.

Further, the screening, according to the reflection spectrum graph, the sample reflection spectrum graph and the sample coal ash content included in each target coal information in the target coal information set, target coal information meeting similar conditions from the target coal information set as similar coal information includes:

classifying the target coal information in the target coal information set according to sample coal ash content included in each target coal information in the target coal information set to obtain a target coal information category set;

determining any one target coal information in each target coal information category in the target coal information category set as the coal representative information corresponding to the target coal information category to obtain a coal representative information set;

determining the sensitivity corresponding to each wavelength index in a preset wavelength index set according to the quantity of the representative coal information in the representative coal information set and a sample reflection spectrum curve graph included in each representative coal information in the representative coal information set;

determining initial similarity corresponding to the coal representative information according to the reflection spectrum curve graph and a sample reflection spectrum curve graph included in each piece of coal representative information in the coal representative information set;

for each piece of coal representative information in the coal representative information set, determining the overall similarity corresponding to the coal representative information according to the sensitivity corresponding to each wavelength index in the wavelength index set, a sample reflection spectrum curve graph included in the coal representative information, the reflection spectrum curve graph and the initial similarity corresponding to the coal representative information;

and screening out corresponding coal representative information with the maximum overall similarity from the coal representative information set to serve as similar coal information.

Further, the determining the overall similarity corresponding to the coal representative information according to the sensitivity corresponding to each wavelength index in the wavelength index set, the sample reflection spectrum graph included in the coal representative information, the reflection spectrum graph, and the initial similarity corresponding to the coal representative information includes:

according to the wavelength index set, screening out a coordinate set from a sample reflection spectrum curve chart included in the coal representative information, wherein the coordinate set is used as a target coordinate set corresponding to the coal representative information;

screening out a spectrum coordinate set from the reflection spectrum curve graph according to the wavelength index set;

and determining the overall similarity corresponding to the coal representative information according to the sensitivity corresponding to each wavelength index in the wavelength index set, the spectrum coordinate set, the target coordinate set corresponding to the coal representative information and the initial similarity.

Further, the generating target coal quality information corresponding to the coal to be detected according to the target coal ash content includes:

when the target coal ash content is larger than a preset ash content threshold value, generating target coal quality information representing that the coal quality to be detected is unqualified;

and when the target coal ash content is less than or equal to an ash content threshold value, generating target coal quality information representing that the quality of the coal to be detected is qualified.

The invention has the following beneficial effects:

the invention relates to coal quality based on image dataThe quantity detection method solves the technical problems of low efficiency and accuracy of coal quality detection by image processing of the target surface image and the reflection spectrum curve chart, and improves the efficiency and accuracy of coal quality detection. Firstly, acquiring a target surface image and a reflection spectrum curve chart of coal to be detected. In practical situations, when detecting the quality of coal, the following methods are generally adopted: and judging the coal quality according to the coal ash content of the coal. The coal ash content of coal is often detected by a burning method, which is a coal ash content detection method widely applied. However, when the coal ash component of coal is detected by using a burning method, the implementation process is often complex, the analysis period is often long, the efficiency of detecting the coal quality is often low, the detection result is often influenced by manual operation, and when the manual operation is not proper, the detection result obtained by detecting the coal quality is often inaccurate, so that the accuracy of detecting the coal quality is low. Secondly, the on-line coal ash content detection method is mainly a radiation method and depends on rays for detection. Among others, radiation methods may include, but are not limited to: low energy

Radiation anti-scatter method, dual energy

Ray casting method and natural

Radiation method. Although the online coal ash content detection method reduces the influence caused by human factors to a certain extent, the radiation method often causes potential safety hazards to workers and the surrounding environment, and the used waste radioactive source is not easy to treat. Therefore, the target surface image and the reflection spectrum curve chart are subjected to image processing, so that the coal quality is detected, and the detection result can be prevented from being influenced by manual operation and damaged by radiation. Then, the target surface image is subjected to adaptive threshold edge detectionAnd obtaining a target texture characteristic and a coal particle outline set. In practice, the coal ash content of the coal to be detected is often related to the size of the coal particles. Therefore, the self-adaptive threshold edge detection is carried out on the target surface image, so that the coal particle profile corresponding to the obtained coal particles is more accurate, the coal ash content of the coal to be detected can be conveniently determined subsequently, and the accuracy of determining the coal ash content of the coal to be detected subsequently can be improved. Secondly, the target texture characteristics can represent the texture condition of the surface of the coal to be detected, and the subsequent detection of the quality of the coal to be detected can be facilitated. And then, determining the coal granularity corresponding to the coal to be detected according to the target texture characteristics and the coal particle contour set. In actual conditions, compared with the method only considering the average gray value and comprehensively considering the target texture characteristics and the coal particle contour set, the method can better reflect the texture condition and the coal particle size of the coal to be detected, and therefore the accuracy of detecting the coal ash content of the coal to be detected subsequently can be improved. And continuing, screening sample coal information meeting the particle condition from a pre-obtained sample coal information set according to the coal granularity, and obtaining a target coal information set as target coal information, wherein the sample coal information in the sample coal information set comprises: the particle condition is that the sample coal particle size included in the sample coal information is equal to the coal particle size. In actual conditions, the target coal information set is screened from the sample coal information set, so that the target coal information set can be conveniently analyzed subsequently, and compared with the method for directly analyzing the sample coal information set, the method has the advantages that the calculated amount is reduced, the occupation of calculation resources is reduced, and the efficiency of detecting the coal quality to be detected is improved. And then, screening target coal information meeting similar conditions from the target coal information set according to the reflection spectrum curve graph, the sample reflection spectrum curve graph and the sample coal ash content included in each target coal information in the target coal information set, and taking the target coal information as similar coal information. Since the target coal information includes sample coal ash, from the set of target coal informationSimilar coal information is screened out, and the coal ash content of the coal to be detected can be indirectly determined. And then, determining the sample coal ash content included in the similar coal information as the target coal ash content of the coal to be detected. In practical cases, since the similar coal information is the target coal information satisfying the similar condition, the target coal ash may be approximately equal to the sample coal ash included in the similar coal information, where the target coal ash may be the coal ash of the coal to be detected. And finally, generating target coal quality information corresponding to the coal to be detected according to the target coal ash content. Therefore, the invention solves the technical problems of low efficiency and accuracy of coal quality detection by processing the target surface image and the reflection spectrum curve chart, and improves the efficiency and accuracy of coal quality detection.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.

Fig. 1 is a flowchart of a coal quality detection method based on image data according to the present invention.

Detailed Description

To further explain the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of the embodiments, structures, features and effects of the technical solutions according to the present invention will be given with reference to the accompanying drawings and preferred embodiments. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

The invention provides a coal quality detection method based on image data, which comprises the following steps:

acquiring a target surface image and a reflection spectrum curve chart of coal to be detected;

performing self-adaptive threshold edge detection on the target surface image to obtain a target texture feature and a coal particle contour set;

determining the coal granularity corresponding to the coal to be detected according to the target texture characteristics and the coal particle outline set;

according to the coal granularity, screening sample coal information meeting the particle condition from a pre-obtained sample coal information set, and taking the sample coal information as target coal information to obtain a target coal information set;

screening target coal information meeting similar conditions from the target coal information set according to the reflection spectrum curve graph, the sample reflection spectrum curve graph and the sample coal ash content included in each target coal information in the target coal information set, and taking the target coal information as similar coal information;

determining sample coal ash content included in the similar coal information as target coal ash content of the coal to be detected;

and generating target coal quality information corresponding to the coal to be detected according to the target coal ash content.

The following steps are detailed:

referring to FIG. 1, a flow diagram of some embodiments of a method for coal quality detection based on image data is shown, in accordance with the present invention. The coal quality detection method based on the image data comprises the following steps:

s1, acquiring a target surface image and a reflection spectrum curve chart of coal to be detected.

In some embodiments, an image of the target surface of the coal to be detected and a reflectance spectrum plot may be obtained.

Wherein the coal to be detected can be the coal of which the quality condition is to be detected. The target surface image may be an image of the surface of the coal to be detected after preprocessing. Pre-processing may include, but is not limited to: graying, denoising and image enhancement. The reflection spectrum curve is also called as a reflection spectrum curve. The reflection spectrum curve chart can represent the curve of the change rule of the spectrum reflectivity of the coal to be detected along with the wavelength. The abscissa of the reflection spectrum graph may be wavelength and the ordinate may be reflectance. As an example, this step may include the steps of:

firstly, acquiring a coal surface image of coal to be detected.

The coal surface image may be an image of the surface of coal to be detected.

For example, an image of the coal surface may be acquired by a camera.

And secondly, graying and denoising the coal surface image to obtain a target surface image.

And thirdly, acquiring a reflection spectrum curve graph.

For example, a reflectance spectrum plot is obtained by a miniature fiber optic spectrometer.

And S2, carrying out self-adaptive threshold edge detection on the target surface image to obtain a target texture feature and a coal particle contour set.

In some embodiments, adaptive threshold edge detection may be performed on the target surface image to obtain a target texture feature and a coal particle contour set.

The target texture characteristics can represent the texture condition of the surface of the coal to be detected. The coal particle contour in the coal particle contour set can be a closed edge representing coal particles contained on the surface of coal to be detected, which is obtained through adaptive threshold edge detection. The adaptive threshold edge detection may be a dual threshold detection. The coal particles may be particles on the surface of coal.

As an example, this step may include the steps of:

firstly, determining a target gradient histogram corresponding to the target surface image according to the target surface image.

Wherein the target gradient histogram may be a gradient histogram of the target surface image.

This step can be implemented by the prior art, and is not described herein again.

And secondly, determining a high threshold value of the self-adaptive threshold value edge detection according to the target gradient histogram.

Wherein the high threshold may be the higher of the dual thresholds included in the adaptive threshold edge detection.

For example, the high threshold for adaptive threshold edge detection is determined by Otsu thresholding from the target gradient histogram.

And thirdly, determining a value of a preset value which is different from the high threshold value by a preset value as a low threshold value of the self-adaptive threshold value edge detection.

The preset value may be a preset value. For example, the preset value may be 1. The low threshold may be the lower of the dual thresholds included in the adaptive threshold edge detection.

In practical situations, the high threshold is usually greater than 0, and the low threshold obtained by subtracting the preset value from the high threshold is usually greater than 0.

And fourthly, performing edge detection on the target surface image according to the high threshold and the low threshold to obtain an edge image.

Wherein the edge image may include: and (5) edge collection. The edges in the edge set may be edges in an edge image.

For example, the edge image may be obtained by performing edge detection on the target surface image through dual-threshold detection, where two thresholds of the dual-threshold detection may be a high threshold and a low threshold, respectively. The edges in the edge set may not include the boundaries of the photographed coal to be detected.

And fifthly, determining reference texture features according to the edge images.

For example, this step may include the following sub-steps:

the first sub-step, according to the edge image, determining the reference gray level co-occurrence matrix.

Wherein, the reference gray level co-occurrence matrix may be a gray level co-occurrence matrix of the edge image.

And a second sub-step of determining energy and inverse variance according to the reference gray level co-occurrence matrix.

The energy can represent the gray level distribution uniformity and texture thickness of the edge image. The inverse variance may characterize the magnitude of local changes in texture of the edge image.

And a third sub-step of determining a reference texture feature according to the energy and the inverse variance.

For example, the formula for determining the reference texture feature may be:

wherein the content of the first and second substances,Qis a reference texture feature.exp() Is an exponential function with a natural constant as the base.ASMIs energy.IDMIs the inverse variance.

In practical situations, the target surface image is subjected to adaptive threshold edge detection, mainly to detect coal particles, so that the subsequent determination of the coal particle size is facilitated. Because the shot coal particle boundary is usually fine, the reference texture features can represent the texture condition of the surface of the coal to be detected after edge detection is carried out, and therefore, the larger the reference texture features are, the finer the texture corresponding to the edge image is, the more likely the coal particle boundary is to be detected, and the better the edge detection effect is. The energy can represent the uniformity degree of gray level distribution and the texture thickness degree of the edge image. If the element values in the gray level co-occurrence matrix are similar, the energy is obtainedASMOften smaller, often indicating fine texture. If the element values in the gray level co-occurrence matrix are not similar, the energy isASMOften large, and often indicative of a less detailed texture. The inverse variance may characterize the magnitude of local changes in texture of the edge image. If the texture between different regions in the edge image is uniform, the variation is slow, and the inverse variance is largeIDMTend to be larger and vice versa smaller. Therefore, when energy is appliedASMThe smaller, the inverse varianceIDMThe larger the reference texture feature is, the finer the texture corresponding to the edge image is, and the better the edge detection effect at this time is. And is

The value range of the reference texture features can be made to be [0,1 ]]Subsequent processing can be facilitated. The formula for determining the reference texture features is not limited to the above formula, and as long as the formula conforms to the above rules, the formula can be used as the formula for determining the reference texture features.

And sixthly, determining fine edge characteristics according to the target surface image and the edge set.

Wherein the fine edge feature may characterize a relatively fine edge condition on the target surface image.

For example, this step may include the following sub-steps:

the first substep, according to the above-mentioned target surface image, confirms the histogram of the gray level.

The gray histogram may be a gray histogram of the target surface image.

A second substep of performing threshold segmentation on the gray level histogram to obtain a gray level threshold.

The gray threshold value can represent pixel points forming a dark area in the target surface image, and the allowed maximum gray value is obtained. The dark region may be a region composed of pixel points in the target surface image having a gray value less than or equal to a gray threshold.

For example, the gray level threshold may be obtained by the Otsu threshold method according to the gray level histogram.

And a third substep of dividing the target surface image according to a gray threshold value to obtain a highlight area set.

The highlight areas in the highlight area set may be areas formed by pixel points in the target surface image, where the gray values of the pixel points are greater than the gray threshold.

In practical situations, due to unevenness, the coal particles are often brighter than the areas except for the coal particles in the coal to be detected in the captured image, and therefore, the divided high-brightness areas are often the captured coal particles.

And a fourth substep, screening out a fine edge set from the edge set according to the highlight region set.

Wherein the thin edges in the thin edge set may be edges inside the highlight region.

And a fifth substep of determining a fine edge feature according to the highlight region set and the fine edge set.

For example, the formula for determining the fine edge feature may be:

wherein, the first and the second end of the pipe are connected with each other,Wis a fine edge feature.LIs the number of pixel points in the target surface image.

Is the first in the tiny edge setiThe number of pixels in each thin edge.iIs the number of the thin edges in the thin edge set.pIs the number of thin edges in the thin edge set.

In practical situations, the target surface image is subjected to adaptive threshold edge detection, mainly to detect coal particles, so that the subsequent determination of the coal particle size is facilitated. Because the shot coal particles often have finer textures, i.e., fine edges, than the coal particle boundaries. These fine edges tend to interfere with subsequent determinations of coal particle size. Due to the fact that

The ratio of the number of the pixel points in the fine edge to the number of the pixel points in the target surface image is obtained, so that the fine edge feature is obtained when the ratio of the number of the pixel points in the fine edge to the number of the pixel points in the target surface image is largerWThe larger the size, the worse the edge detection effect at this time. The formula for determining the fine edge feature is not limited to the above formula, and as long as the formula conforms to the above rules, the formula for determining the fine edge feature can be used.

And seventhly, determining the complete characteristics of the edge according to the edge set.

The edge integrity feature can characterize the integrity of the edge in the edge set.

For example, this step may include the following sub-steps:

the first substep, divide the edge in the edge set, get the edge set of closing and not closing the edge set.

Wherein the closure edges in the closure edge set may be end-to-end edges. The non-closed edges in the set of non-closed edges may be edges in the set of edges other than the set of closed edges.

And in the second sub-step, determining the edge integrity characteristic according to the number of the non-closed edges in the non-closed edge set, the number and the area of the closed edges in the closed edge set and the average value of the areas of the closed edges in the closed edge set.

For example, the formula for determining the edge integrity feature may be:

wherein the content of the first and second substances,Eis an edge integrity feature.uThe number of non-closed edges in the set of non-closed edges.JIs the number of closed edges in the set of closed edges.

Is the first in the closed edge setjArea of each closed edge. For example, the number of pixels in the closed edge may characterize the area of the closed edge.jIs the sequence number of the closure edge in the closure edge set.sIs the average of the areas of the closed edges in the set of closed edges.

In practical situations, the target surface image is subjected to adaptive threshold edge detection, mainly to detect coal particles, so that the subsequent determination of the coal particle size is facilitated. Since the boundaries of coal particles tend to be closed and the size of individual coal particles tends to be similar. Therefore, the more non-closed edges are detected, the worse the edge detection effect is, and the worse the similarity of the sizes of the coal particles is, the worse the edge detection effect isThe less good. Thus, the number of non-closure edges in the set of non-closure edgesuThe more or

The larger the size, the larger the edge integrity feature is, and the worse the edge detection effect at this time is. The formula for determining the complete edge feature is not limited to the above formula, and as long as the formula conforms to the above rule, the formula can be used as the formula for determining the complete edge feature.

And eighthly, determining an edge detection effect index according to the reference texture feature, the fine edge feature and the edge integrity feature.

For example, the formula for determining the edge detection effect index may be:

wherein the content of the first and second substances,Ris an edge detection effect index.WIs a fine edge feature.EIs an edge integrity feature.QIs a reference texture feature.exp() Is an exponential function with a natural constant as the base.

In practice, the texture feature is referred toQLarger, fine edge featuresWOr edge integrity featureEThe smaller the edge detection effect indexRThe larger the size, the better the edge detection effect at this time. And is

Can make the edge detection effect indexRHas a value range of [0,1]The subsequent processing can be facilitated. The formula for determining the edge detection effect index is not limited to the above formula, and as long as the formula conforms to the above rule, the formula can be used as the formula for determining the edge detection effect index.

And ninthly, determining a target low threshold according to the edge detection effect index and a preset effect index threshold.

The effect index threshold may be a preset maximum edge detection effect index that is allowed when the low threshold needs to be updated. For example, the effectiveness index threshold may be 0.7.

For example, this step may include the following sub-steps:

a first substep of determining a low threshold as a target low threshold when the edge detection effect indicator is greater than the effect indicator threshold.

And a second substep of updating the low threshold to the low threshold minus a preset value when the low threshold is greater than 0 and the edge detection effect index is less than or equal to the effect index threshold.

And a third substep of performing edge detection on the target surface image according to the high threshold and the updated low threshold to obtain an updated edge image.

Wherein updating the edge image may include: and updating the edge set.

And a fourth substep of determining an updated reference texture feature based on the updated edge image.

And a fifth substep of determining updated fine edge features based on the target surface image and the updated edge set.

And a sixth substep of determining an updated edge integrity feature based on the updated edge set.

And a seventh substep of determining an updated edge detection effect index according to the updated reference texture feature, the updated fine edge feature and the updated edge integrity feature.

The specific implementation manners of the third to seventh substeps included in the ninth step included in step S2 may refer to the fourth to eighth steps included in step S2, and the updated low threshold may be used as the low threshold, and the fourth to eighth steps included in step S2 are executed to obtain the edge image, the edge set, the reference texture feature, the fine edge feature, the edge integrity feature, and the edge detection effect index, that is, the updated edge image, the updated edge set, the updated reference texture feature, the updated fine edge feature, the updated edge integrity feature, and the updated edge detection effect index.

And a seventh substep of, when the updated low threshold is greater than 0 and the updated edge detection effect index is less than or equal to the effect index threshold, taking the updated low threshold as the low threshold, repeating the above steps until the updated edge detection effect index is greater than the effect index threshold or the updated low threshold is not greater than 0, and stopping repeating.

Wherein, the above steps may include: the ninth step included in step S2 includes the second to sixth substeps.

And an eighth substep of determining the updated low threshold as the target low threshold when the updated edge detection effect indicator is greater than the effect indicator threshold.

And a ninth substep of determining the edge detection effect index and the maximum value of all the updated edge detection effect indexes as a target detection effect index and determining a low threshold corresponding to the target detection effect index as a target low threshold when the updated low threshold is not greater than 0.

The low threshold corresponding to the target detection effect index may be a low threshold in determining the target detection effect index. For example, according to the edge set, the determined low threshold corresponding to the edge detection effect index may be a low threshold when the edge set is obtained. According to the updated edge set, the determined low threshold corresponding to the updated edge detection effect index may be an updated low threshold when the updated edge set is obtained.

For example, when the updated low threshold is not greater than 0, the edge detection effect index may be 0.3, and all the updated edge detection effect indexes may be: 0.5, 0.4 and 0.6. The maximum value of the edge detection effect index and all the updated edge detection effect indexes may be 0.6.

And step ten, determining the reference texture feature corresponding to the target low threshold as the target texture feature.

The reference texture feature corresponding to the target low threshold may be a reference texture feature determined according to the target edge image. The target edge image may be an edge image obtained by performing edge detection on the target surface image according to the target low threshold and the target high threshold.

And step ten, carrying out edge detection on the target surface image according to the high threshold and the target low threshold to obtain a coal particle outline set.

Wherein the coal particle profile in the coal particle profile set can represent the boundary of the shot coal particle.

And S3, determining the coal granularity corresponding to the coal to be detected according to the target textural features and the coal particle outline set.

In some embodiments, the coal particle size corresponding to the coal to be detected may be determined according to the target texture feature and the coal particle contour set.

Wherein, the coal granularity can be the granularity of the coal to be detected.

As an example, this step may include the steps of:

firstly, determining the area average value of the coal particle outline according to the coal particle outline set.

The coal particle profile area average value can be an average value of areas of coal particle profiles in the coal particle profile set. The area of the coal particle profile can be characterized by the number of pixel points in the coal particle profile.

And secondly, determining the coal granularity according to the coal particle contour area average value and the target texture characteristics.

For example, the formula for determining the coal particle size may be:

wherein the content of the first and second substances,Tis the coal particle size.

Is the target texture feature.sIs the average of the areas of the closed edges in the set of closed edges.

In practical situations, when the target texture features

Or in a closed edge setMean of area of closed edgessThe larger the coal particle sizeTThe larger the tendency.

And S4, screening sample coal information meeting the particle condition from a pre-obtained sample coal information set according to the coal granularity, and taking the sample coal information as target coal information to obtain a target coal information set.

In some embodiments, sample coal information meeting the particle condition may be screened from a pre-obtained sample coal information set according to the coal particle size, and the sample coal information is used as target coal information to obtain a target coal information set.

The sample coal information in the sample coal information set may include: sample coal ash, sample reflectance spectrum profile, and sample coal particle size. The particle condition may be that the sample coal information includes a sample coal particle size equal to the above-described coal particle size. The sample coal ash can be a coal ash of the sample coal. The shape and size of the sample coal may be the same as the shape and size of the coal to be detected. The sample coal may be coal of known coal ash content. The sample coal may correspond to the sample coal information one to one. The sample reflection spectrum curve graph can represent the curve of the spectral reflectivity of the sample coal along with the change rule of the wavelength. The abscissa of the plot of the sample reflectance spectrum may be wavelength and the ordinate may be reflectance. The sample coal particle size may be the particle size of the sample coal.

As an example, obtaining the sample reflection spectrum graph and the sample coal granularity included in the sample coal information set may refer to the above-described manner of obtaining the reflection spectrum graph and the coal granularity.

In practical situations, when the number of sample coal information in the sample coal information set is larger, the types of sample coal are more, and the coal ash content of the coal to be detected determined subsequently is more accurate.

And S5, screening target coal information meeting similar conditions from the target coal information set according to the reflection spectrum curve graph, the sample reflection spectrum curve graph and the sample coal ash content included in each target coal information in the target coal information set, and taking the target coal information as similar coal information.

In some embodiments, target coal information satisfying similar conditions may be screened from the target coal information set according to the reflection spectrum graph, a sample reflection spectrum graph included in each target coal information in the target coal information set, and a sample coal ash content, and may be used as similar coal information.

Wherein the similarity condition can be that the sample coal is similar to the coal to be detected.

As an example, this step may comprise the steps of:

the method comprises the steps of firstly, classifying target coal information in the target coal information set according to sample coal ash content included in each target coal information in the target coal information set to obtain a target coal information category set.

Wherein, the target coal information in the target coal information category comprises the same sample coal ash content.

For example, the target coal information having the same sample coal ash content included in the target coal information set may be classified into the same target coal information category.

And secondly, determining any one target coal information in each target coal information category in the target coal information category set as the coal representative information corresponding to the target coal information category to obtain a coal representative information set.

And thirdly, determining the sensitivity corresponding to each wavelength index in a preset wavelength index set according to the quantity of the representative coal information in the representative coal information set and a sample reflection spectrum curve graph included in each representative coal information in the representative coal information set.

Wherein, the wavelength index in the wavelength index set can be a preset wavelength. For example, the set of wavelength indicators may be: [400 nm, 500 nm, 600 nm, 700 nm, 800 nm, 900 nm, 1000 nm, 1100 nm ].

For example, the formula for determining the sensitivity correspondence corresponding to each wavelength index in the preset wavelength index set may be:

wherein the content of the first and second substances,

is the first in the wavelength index setkThe sensitivity corresponding to each wavelength index.kIs the number of the wavelength index in the wavelength index set.NIs the amount of coal representative information in the coal representative information set.

Is the first in the coal representative information setnAnd the ordinate is included in the coordinate corresponding to the target coordinate point in the sample reflection spectrum curve chart included in the coal representing information, wherein the ordinate is the reflectivity. The coordinates corresponding to the target coordinate point may include an abscissa that may be associated with the second coordinate pointkThe wavelength indexes are equal.nThe number is the serial number of the coal representative information in the coal representative information set.

In practical situations, under the same wavelength index, the larger the influence of the change of the reflectivity on the sample coal ash content is, the more sensitive the sample coal ash content is under the wavelength index is, and the more the reflectivity can be used for representing the sample coal ash content. As a result of this, it is possible to,

can be characterized inkUnder the index of each wavelength, the reflectivity is influenced by the change of the ash content of the sample coal, so that,

can characterize the first in the wavelength index setkThe sensitivity corresponding to each wavelength index. The formula for determining the sensitivity corresponding to the wavelength index is not limited to the above formula, and the formula conforming to the above rule may be used as the formula for determining the sensitivity corresponding to the wavelength index.

And fourthly, determining the initial similarity corresponding to the coal representative information according to the reflection spectrum graph and a sample reflection spectrum graph included in each piece of coal representative information in the coal representative information set.

The initial similarity corresponding to the coal representative information can represent the similar situation between the sample reflection spectrum graph and the reflection spectrum graph included in the coal representative information.

For example, the profile similarity between the sample reflectance spectrum graph and the reflectance spectrum graph included in the coal representative information may be determined by the shape context algorithm as the initial similarity corresponding to the coal representative information.

And fifthly, for each piece of coal representative information in the coal representative information set, determining the overall similarity corresponding to the coal representative information according to the sensitivity corresponding to each wavelength index in the wavelength index set, the sample reflection spectrum curve graph included in the coal representative information, the reflection spectrum curve graph and the initial similarity corresponding to the coal representative information.

The overall similarity corresponding to the coal representation information can represent the similarity between the sample coal and the coal to be detected.

For example, this step may include the following sub-steps:

and a first substep of screening out a coordinate set from a sample reflection spectrum curve chart included in the coal representative information according to the wavelength index set, wherein the coordinate set is used as a target coordinate set corresponding to the coal representative information.

The abscissa included in the target coordinate set corresponding to the coal representation information may be equal to the wavelength index in the wavelength index set. The target coordinates in the target coordinate set may correspond to the wavelength indexes in the wavelength index set one to one.

And a second substep of screening out a spectrum coordinate set from the reflection spectrum curve graph according to the wavelength index set.

The spectrum coordinates in the spectrum coordinate set may include an abscissa equal to the wavelength index in the wavelength index set. The spectral coordinates in the set of spectral coordinates may correspond one-to-one to the wavelength indices in the set of wavelength indices.

And a third substep of determining the overall similarity corresponding to the coal representative information according to the sensitivity corresponding to each wavelength index in the wavelength index set, the spectrum coordinate set, the target coordinate set corresponding to the coal representative information and the initial similarity.

For example, the formula for determining the overall similarity corresponding to the coal representation information may be:

wherein, the first and the second end of the pipe are connected with each other,

is the first in the coal representative information setnAnd the coal represents the overall similarity corresponding to the information.

Is the first in the coal representative information setnAnd representing the initial similarity corresponding to the information by the coal.

Is the first in the wavelength index setkThe sensitivity corresponding to each wavelength index.kIs the number of the wavelength index in the wavelength index set.KIs the number of wavelength indices in the set of wavelength indices.

Is the ordinate comprised by the target spectral coordinates. The target spectral coordinate may be the abscissa and the second coordinate included in the set of spectral coordinateskThe individual wavelength indexes equal spectral coordinates.

Is the ordinate comprised by the representative coordinate. The representative coordinates may benIncluded in a target coordinate set corresponding to each coal representative informationAbscissa and ordinate ofkThe wavelength indexes are equal target coordinates.

Is a preset value greater than 0.

The denominator is 0 and tends to be small.

In practical cases, when coal represents the first in the information setnInitial similarity corresponding to individual coal representative information

The larger, thenThe more similar the sample coal corresponding to the individual coal representative information is to the coal to be detected. When the wavelength index is the first in the setkSensitivity corresponding to each wavelength index

The larger the ash content, the coal ash content tends to be indicated inkThe more sensitive the wavelength index is, the more the reflectivity can be used for representing the ash content of the sample coal, and at the moment, the more credible the similarity between the sample coal and the coal to be detected is represented by adopting the absolute value of the reflectivity difference between the sample coal and the coal to be detected. That is, the smaller the absolute value of the reflectivity difference between the sample coal and the coal to be detected is, the more similar the sample coal and the coal to be detected is. The formula for determining the overall similarity corresponding to the coal representative information is not limited to the above formula, and as long as the formula conforms to the above rule, the formula can be used as the formula for determining the overall similarity corresponding to the coal representative information.

And sixthly, screening out the corresponding coal representative information with the maximum overall similarity from the coal representative information set to serve as similar coal information.

And S6, determining the sample coal ash content included by the similar coal information as the target coal ash content of the coal to be detected.

In some embodiments, the sample coal ash included in the similar coal information may be determined as the target coal ash of the coal to be detected.

Wherein the target coal ash can be the coal ash of the coal to be detected.

And S7, generating target coal quality information corresponding to the coal to be detected according to the target coal ash content.

In some embodiments, target coal quality information corresponding to the coal to be detected may be generated according to the target coal ash.

The target coal quality information can represent the quality condition of the coal to be detected.

As an example, this step may include the steps of:

step one, when the target coal ash content is larger than a preset ash content threshold value, target coal quality information representing that the coal quality to be detected is unqualified is generated.

The ash threshold value can be the maximum allowable target coal ash when the preset quality of the coal to be detected is qualified. For example, the ash threshold may be 12%. The unqualified quality of the coal to be detected can represent that the quality of the coal to be detected does not meet the production standard. The qualified quality of the coal to be detected can represent that the quality of the coal to be detected meets the production standard.

And secondly, generating target coal quality information representing that the quality of the coal to be detected is qualified when the target coal ash content is less than or equal to an ash content threshold value.

According to the coal quality detection method based on the image data, the technical problem that the efficiency and the accuracy of detecting the coal quality are low is solved by carrying out image processing on the target surface image and the reflection spectrum curve chart, and the efficiency and the accuracy of detecting the coal quality are improved. Firstly, acquiring a target surface image and a reflection spectrum curve chart of coal to be detected. In practical situations, when detecting the coal quality, the following methods are generally adopted: and judging the coal quality according to the coal ash content of the coal. The coal ash content of coal is often detected by a burning method, which is a coal ash content detection method widely applied. However, when coal is detected by the combustion methodWhen the coal ash is used for time sharing, the implementation process is often complex, the analysis period is often long, the efficiency of detecting the coal quality is often low, the detection result is often influenced by manual operation, when the manual operation is improper, the detection result obtained by detecting the coal quality is often inaccurate, and the accuracy of detecting the coal quality is low. Secondly, the on-line coal ash content detection method is mainly a radiation method and depends on rays for detection. Among others, radiation methods may include, but are not limited to: low energy

Radiation anti-scatter method, dual energy

Ray casting method and nature

Radiation method. Although the online coal ash content detection method reduces the influence caused by human factors to a certain extent, the radiation method often causes potential safety hazards to workers and the surrounding environment, and the used waste radioactive source is not easy to treat. Therefore, the target surface image and the reflection spectrum curve chart are subjected to image processing, so that the coal quality is detected, and the detection result can be prevented from being influenced by manual operation and damaged by radiation. And then, carrying out self-adaptive threshold edge detection on the target surface image to obtain a target texture feature and a coal particle outline set. In practical situations, the coal ash content of the coal to be detected is often related to the size of the coal particles. Therefore, the self-adaptive threshold edge detection is carried out on the target surface image, so that the coal particle profile corresponding to the obtained coal particles is more accurate, the coal ash content of the coal to be detected can be conveniently determined subsequently, and the accuracy of determining the coal ash content of the coal to be detected subsequently can be improved. Secondly, the target texture characteristics can represent the texture condition of the surface of the coal to be detected, and the subsequent detection of the quality of the coal to be detected can be facilitated. Then, according to the target texture characteristics and the coal particle outline setAnd determining the coal granularity corresponding to the coal to be detected. In actual conditions, compared with the method that only the average gray value is considered, the target texture characteristic and the coal particle contour set are comprehensively considered, the texture condition and the coal particle size of the coal to be detected can be reflected, and therefore the accuracy of detecting the coal ash content of the coal to be detected subsequently can be improved. And continuing, screening sample coal information meeting the particle condition from a pre-obtained sample coal information set according to the coal granularity, and obtaining a target coal information set as target coal information, wherein the sample coal information in the sample coal information set comprises: the particle condition is that the sample coal particle size included in the sample coal information is equal to the coal particle size. In actual conditions, the target coal information set is screened from the sample coal information set, so that the target coal information set can be conveniently analyzed subsequently, and compared with the method for directly analyzing the sample coal information set, the method has the advantages that the calculated amount is reduced, the occupation of calculation resources is reduced, and the efficiency of detecting the coal quality to be detected is improved. And then, screening target coal information meeting similar conditions from the target coal information set according to the reflection spectrum curve graph, the sample reflection spectrum curve graph and the sample coal ash content included in each target coal information in the target coal information set, and taking the target coal information as similar coal information. Because the target coal information comprises the sample coal ash content, similar coal information is screened out from the target coal information set, and the coal ash content of the coal to be detected can be indirectly determined. And then, determining the sample coal ash content included in the similar coal information as the target coal ash content of the coal to be detected. In practical cases, since the similar coal information is the target coal information satisfying the similar condition, the target coal ash may be approximately equal to the sample coal ash included in the similar coal information, where the target coal ash may be the coal ash of the coal to be detected. And finally, generating target coal quality information corresponding to the coal to be detected according to the target coal ash content. Therefore, the invention carries out image processing on the target surface image and the reflection spectrum curve chartThe technical problem that the efficiency and the accuracy of detecting the coal quality are low is solved, and the efficiency and the accuracy of detecting the coal quality are improved.

The above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications or substitutions do not cause the essential features of the corresponding technical solutions to depart from the scope of the technical solutions of the embodiments of the present application, and are intended to be included within the scope of the present application.

Claims (10)

1. A coal quality detection method based on image data is characterized by comprising the following steps:

acquiring a target surface image and a reflection spectrum curve chart of coal to be detected;

carrying out self-adaptive threshold edge detection on the target surface image to obtain a target texture feature and a coal particle outline set;

determining the coal granularity corresponding to the coal to be detected according to the target texture characteristics and the coal particle outline set;

according to the coal granularity, screening sample coal information meeting the particle condition from a pre-obtained sample coal information set, and obtaining a target coal information set as target coal information, wherein the sample coal information in the sample coal information set comprises: sample coal ash content, a sample reflectance spectrum curve graph and sample coal granularity, wherein the particle condition is that the sample coal granularity included in the sample coal information is equal to the coal granularity;

screening target coal information meeting similar conditions from the target coal information set according to the reflection spectrum curve graph, and a sample reflection spectrum curve graph and sample coal ash content included in each target coal information in the target coal information set, and taking the target coal information as similar coal information;

determining the sample coal ash content included in the similar coal information as the target coal ash content of the coal to be detected;

and generating target coal quality information corresponding to the coal to be detected according to the target coal ash content.

2. The method of claim 1, wherein the performing adaptive threshold edge detection on the target surface image to obtain a target texture feature and a coal particle contour set comprises:

determining a target gradient histogram corresponding to the target surface image according to the target surface image;

determining a high threshold of adaptive threshold edge detection according to the target gradient histogram;

determining a numerical value which is different from the high threshold value by a preset numerical value as a low threshold value of the self-adaptive threshold value edge detection;

according to the high threshold and the low threshold, performing edge detection on the target surface image to obtain an edge image, wherein the edge image comprises: collecting edges;

determining reference texture features according to the edge images;

determining fine edge features according to the target surface image and the edge set;

determining the complete characteristics of the edge according to the edge set;

determining an edge detection effect index according to the reference texture feature, the fine edge feature and the edge integrity feature;

determining a target low threshold according to the edge detection effect index and a preset effect index threshold;

determining the reference texture feature corresponding to the target low threshold as a target texture feature;

and carrying out edge detection on the target surface image according to the high threshold and the target low threshold to obtain a coal particle outline set.

3. The method for detecting coal quality based on image data according to claim 2, wherein the determining a target low threshold according to the edge detection effect index and a preset effect index threshold comprises:

when the edge detection effect index is larger than the effect index threshold value, determining the low threshold value as a target low threshold value;

when the low threshold is greater than 0 and the edge detection effect index is less than or equal to the effect index threshold, updating the low threshold to be the low threshold minus a preset value;

according to the high threshold and the updated low threshold, performing edge detection on the target surface image to obtain an updated edge image, wherein the updating the edge image comprises: updating the edge set;

determining an updated reference texture feature according to the updated edge image;

determining updated tiny edge features according to the target surface image and the updated edge set;

determining the complete characteristics of the update edge according to the update edge set;

determining an updated edge detection effect index according to the updated reference texture feature, the updated fine edge feature and the updated edge integrity feature;

when the updated low threshold is greater than 0 and the updated edge detection effect index is less than or equal to the effect index threshold, taking the updated low threshold as the low threshold, repeating the steps until the updated edge detection effect index is greater than the effect index threshold or the updated low threshold is not greater than 0, and stopping repeating;

when the updated edge detection effect index is larger than the effect index threshold, determining the updated low threshold as a target low threshold;

and when the updated low threshold value is not more than 0, determining the edge detection effect index and the maximum value of all the updated edge detection effect indexes as a target detection effect index, and determining the low threshold value corresponding to the target detection effect index as a target low threshold value.

4. The method of claim 2, wherein the determining the reference texture feature according to the edge image comprises:

determining a reference gray level co-occurrence matrix according to the edge image;

determining energy and inverse variance according to the reference gray level co-occurrence matrix;

and determining the reference texture feature according to the energy and the inverse variance.

5. The method of claim 2, wherein the determining fine edge features according to the target surface image and the edge set comprises:

determining a gray level histogram according to the target surface image;

performing threshold segmentation on the gray level histogram to obtain a gray level threshold;

dividing the target surface image according to a gray threshold value to obtain a highlight area set;

screening out a fine edge set from the edge set according to the highlight area set;

and determining the fine edge characteristics according to the highlight region set and the fine edge set.

6. The method for detecting coal quality based on image data according to claim 2, wherein the determining edge integrity features according to the edge set comprises:

dividing edges in the edge set to obtain a closed edge set and a non-closed edge set;

and determining the edge integrity feature according to the number of the non-closed edges in the non-closed edge set, the number and the area of the closed edges in the closed edge set and the average value of the areas of the closed edges in the closed edge set.

7. The method for detecting coal quality based on image data according to claim 1, wherein the determining the coal granularity corresponding to the coal to be detected according to the target texture features and the coal particle contour set comprises:

determining the area mean value of the coal particle outline according to the coal particle outline set;

and determining the coal granularity according to the coal particle contour area mean value and the target texture feature.

8. The method for detecting coal quality based on image data according to claim 1, wherein the step of screening target coal information satisfying similar conditions from the target coal information set according to the reflection spectrum graph, the sample reflection spectrum graph and the sample coal ash content included in each target coal information in the target coal information set comprises:

classifying the target coal information in the target coal information set according to sample coal ash content included in each target coal information in the target coal information set to obtain a target coal information category set;

determining any one target coal information in each target coal information category in the target coal information category set as the coal representative information corresponding to the target coal information category to obtain a coal representative information set;

determining the sensitivity corresponding to each wavelength index in a preset wavelength index set according to the quantity of the representative coal information in the representative coal information set and a sample reflection spectrum curve graph included in each representative coal information in the representative coal information set;

determining initial similarity corresponding to the coal representative information according to the reflection spectrum curve graph and a sample reflection spectrum curve graph included in each piece of coal representative information in the coal representative information set;

for each piece of coal representative information in the coal representative information set, determining the overall similarity corresponding to the coal representative information according to the sensitivity corresponding to each wavelength index in the wavelength index set, a sample reflection spectrum curve graph included in the coal representative information, the reflection spectrum curve graph and the initial similarity corresponding to the coal representative information;

and screening out corresponding coal representative information with the maximum overall similarity from the coal representative information set to serve as similar coal information.

9. The method of claim 8, wherein the determining the overall similarity corresponding to the representative coal information according to the sensitivity corresponding to each wavelength index in the set of wavelength indexes, the sample reflection spectrum graph included in the representative coal information, the reflection spectrum graph, and the initial similarity corresponding to the representative coal information comprises:

according to the wavelength index set, screening out a coordinate set from a sample reflection spectrum curve chart included in the coal representative information, wherein the coordinate set is used as a target coordinate set corresponding to the coal representative information;

screening out a spectrum coordinate set from the reflection spectrum curve chart according to the wavelength index set;

and determining the overall similarity corresponding to the coal representative information according to the sensitivity corresponding to each wavelength index in the wavelength index set, the spectrum coordinate set, the target coordinate set corresponding to the coal representative information and the initial similarity.

10. The method for detecting coal quality based on image data according to claim 1, wherein the generating target coal quality information corresponding to the coal to be detected according to the target coal ash content comprises:

when the target coal ash content is larger than a preset ash content threshold value, generating target coal quality information representing that the coal quality to be detected is unqualified;

and when the target coal ash content is less than or equal to an ash content threshold value, generating target coal quality information representing that the quality of the coal to be detected is qualified.

Images