CN115082744B - Artificial intelligence-based solar heat collection efficiency analysis method and system - Google Patents
Artificial intelligence-based solar heat collection efficiency analysis method and system Download PDFInfo
- Publication number
- CN115082744B CN115082744B CN202211002558.4A CN202211002558A CN115082744B CN 115082744 B CN115082744 B CN 115082744B CN 202211002558 A CN202211002558 A CN 202211002558A CN 115082744 B CN115082744 B CN 115082744B
- Authority
- CN
- China
- Prior art keywords
- area
- crack
- region
- dirt
- obtaining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 31
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 22
- 230000002159 abnormal effect Effects 0.000 claims abstract description 71
- 238000000034 method Methods 0.000 claims abstract description 31
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 23
- 230000008859 change Effects 0.000 claims description 48
- 239000011159 matrix material Substances 0.000 claims description 44
- 239000002689 soil Substances 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 10
- 230000011218 segmentation Effects 0.000 claims description 9
- 230000007613 environmental effect Effects 0.000 claims description 8
- 239000013598 vector Substances 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 238000012935 Averaging Methods 0.000 claims description 4
- 238000005336 cracking Methods 0.000 claims description 2
- 230000007547 defect Effects 0.000 abstract description 29
- 238000012216 screening Methods 0.000 abstract description 3
- 238000009826 distribution Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 230000004075 alteration Effects 0.000 description 4
- 230000002950 deficient Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000001955 cumulated effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 210000003608 fece Anatomy 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/54—Extraction of image or video features relating to texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, in particular to a solar heat collection efficiency analysis method and system based on artificial intelligence. The method comprises the steps of performing primary judgment according to pixel values of image pixel points to obtain a first abnormal region. And eliminating the color difference area according to the first texture characteristics to obtain a second abnormal area. And obtaining a suspected crack region according to the size and morphology of the second abnormal region, classifying according to the second texture features in the suspected crack region, and screening out the crack region and the dirty region. And analyzing the heat collection efficiency of the solar water heater according to the characteristics of the dirty area and the characteristics of the crack area. The invention classifies the defect types and realizes accurate prediction analysis on the heat collection efficiency aiming at the defect types.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a solar heat collection efficiency analysis method and system based on artificial intelligence.
Background
Solar water heaters are common heating devices, and heat energy in sunlight is absorbed to heat water so as to meet the requirement of people on hot water in life production. A common solar water heater is a vacuum tube water heater. For the region needing regional heat collection, a plurality of groups of solar water heaters are arranged to realize regional heat collection water supply in order to save resources.
Most vacuum tube solar water heater is composed of interval vacuum and back reflection layer, sunlight is reflected into vacuum tube and then absorbed by inner container. The vacuum tube is the initial starting position for energy conversion, so the state of the vacuum tube affects the heat collection situation. Because solar water heaters are often arranged outdoors, various defects can be generated due to environmental influences, such as dirt defects caused by rain stains, bird droppings and the like; or the pipeline has crack defects caused by external force factors; as the vacuum tube works, color difference defects due to aging or other reasons may also occur on the tube. Among the defects, dirt defects and cracks can obviously influence heat collection efficiency, chromatic aberration defects have small influence and even cannot influence heat absorption capacity, the defect area is most commonly identified through images by utilizing a computer vision technology in the existing defect detection, but the chromatic aberration defects and the dirt defects cannot be well distinguished in the analysis process, so that the detection is inaccurate, and the subsequent heat collection analysis is influenced. If a machine learning technique is used, a neural network is used to identify defective pixels, a large amount of defective data is required to train the network, and the detection cost is increased.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a solar heat collection efficiency analysis method and system based on artificial intelligence, and the adopted technical scheme is as follows:
the invention provides a solar heat collection efficiency analysis method based on artificial intelligence, which comprises the following steps:
obtaining a solar pipeline image of a plurality of continuous frames; classifying according to pixel values of pixel points in the solar pipeline image to obtain abnormal pixel points and normal pixel points; obtaining a first abnormal region according to the abnormal pixel points;
obtaining a first gray level co-occurrence matrix of the first abnormal region; obtaining a first texture feature according to the first gray level co-occurrence matrix; removing the first abnormal region corresponding to the first texture feature similar to the standard pipeline texture feature to obtain a second abnormal region;
identifying a strip-shaped abnormal region according to the size of the second abnormal region, wherein the strip-shaped abnormal region is taken as a suspected crack region, and the other second abnormal regions are dirty regions; obtaining a second gray level co-occurrence matrix of each pixel point of the suspected crack region in a preset neighborhood range, and obtaining a second texture feature set of each suspected crack region according to the second gray level co-occurrence matrix; taking the suspected crack areas corresponding to the second texture feature sets with the same similarity as a class of area categories; taking the area category with the minimum data amount as a crack area, and the other areas are the dirty areas;
obtaining color information of the dirt area, and obtaining the dirt thickness according to the color information; obtaining an overall soil characteristic according to the soil thickness and the area of the soil region; obtaining crack change characteristics according to the size change and the pixel value change of the crack region in the continuous multi-frame solar pipeline image; taking the crack change characteristics and the area of the crack area as integral crack characteristics; and predicting solar heat collection efficiency according to the integral dirt characteristics, the integral crack characteristics and the current environmental data.
Further, the obtaining the solar pipeline images of the continuous multiframe includes:
collecting images of the solar water heater of continuous multiframes; sending the solar water heater image into a pre-trained pipeline segmentation network to obtain the solar pipeline image; the solar duct image includes a front duct image and a back duct image.
Further, the obtaining the first texture feature according to the first gray level co-occurrence matrix includes:
and taking the energy, contrast, local relativity and entropy of the first gray level co-occurrence matrix as the first texture features.
Further, the removing the first anomaly region corresponding to the first texture feature that is similar to the standard pipeline texture feature includes:
obtaining a second norm difference between the first texture feature and the standard pipeline texture feature, and taking the inverse of the second norm difference as a judgment index; and removing the first abnormal region corresponding to the judgment index larger than a preset judgment index threshold value.
Further, the obtaining a second texture feature set for each of the suspected crack regions according to the second gray level co-occurrence matrix includes:
converting the second gray level co-occurrence matrix into a one-dimensional space to obtain a one-dimensional feature vector; and taking the set of the one-dimensional feature vectors of all pixel points in the suspected crack area as the second texture feature set.
Further, the obtaining the color information of the dirty region, and obtaining the dirty thickness according to the color information includes:
converting the dirty region into an HIS color space, and taking saturation and brightness information in the HIS color space as the color information;
obtaining the dirt thickness of each dirt area according to a dirt thickness calculation formula; the dirt thickness calculation formula comprises:
wherein ,for the thickness of the soil to be mentioned,for the purpose of the brightness information as described,is the saturation information;
further, the obtaining the soil characteristics from the soil thickness and the area of the soil region includes:
acquiring a front dirty region and a back dirty region according to the front pipeline image and the back pipeline image; the area of the front dirty area and the dirty thickness are respectively adjusted according to a preset front dirty weight; respectively adjusting the area of the back dirt area and the dirt thickness according to preset back dirt weight;
accumulating the adjusted area of the front dirt area and the area of the back dirt area to obtain overall dirt area information; accumulating and averaging the adjusted dirt thickness of the front dirt area and the dirt thickness of the back dirt area to obtain overall dirt thickness information; and taking the whole dirt area information and the whole dirt thickness information as the whole dirt characteristics.
Further, the obtaining the crack variation feature according to the size variation and the pixel value variation of the crack region in the continuous multi-frame solar pipeline image comprises:
obtaining the size change according to a size change formula; the size change formula comprises:
wherein ,is the firstThe size of the dimensions of each of the crack regions varies,is the firstThe area of the crack region in the solar pipe image is framed,a frame number for the solar pipeline image;
obtaining the pixel value change according to a pixel value change formula; the pixel value size variation formula includes:
wherein ,is the firstThe pixel values of each of the crack regions vary in size,is the firstThe average pixel value of the crack region in the solar pipeline image is framed,a frame number for the solar pipeline image;
the crack change feature is a product of the dimensional change and the pixel value change.
Further, the predicting solar heat collection efficiency based on the fouling characteristics, the crack characteristics, and the current environmental data comprises:
and inputting the dirt characteristics, the crack characteristics and the current environment data into a pre-trained heat collection efficiency prediction network, and outputting the solar heat collection efficiency.
The invention also provides an artificial intelligence-based solar heat collection efficiency analysis system, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, and is characterized in that the processor realizes the steps of any one of the artificial intelligence-based solar heat collection efficiency analysis methods when executing the computer program.
The invention has the following beneficial effects:
1. according to the embodiment of the invention, defective pixel points are distinguished according to the difference of pixel values, and a first abnormal region is obtained. Because texture information of the color difference area is not changed, a first abnormal area corresponding to a first texture feature similar to the texture feature of the standard pipeline is removed, the color difference area is distinguished, and a second abnormal area comprising a dirty area and a crack area is obtained. And primarily dividing the crack areas according to the forms of the dirty areas and the crack areas to obtain suspected crack areas, further considering the distribution quantity of the crack areas, and screening the crack areas to realize the classification of defects.
2. According to the embodiment of the invention, the dirty characteristic is obtained according to the dirty color information and the dirty area, the crack characteristic is obtained according to the combination of the size change of the crack and the size change of the pixel value, and the influence of the defect on the solar heat collection efficiency is effectively analyzed according to the dirty characteristic and the crack characteristic.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a solar heat collection efficiency analysis method based on artificial intelligence according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of a specific implementation, structure, characteristics and effects of the artificial intelligence-based solar heat collection efficiency analysis method and system according to the invention with reference to the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
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 solar heat collection efficiency analysis method and a solar heat collection efficiency analysis system based on artificial intelligence, which are specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a solar heat collection efficiency analysis method based on artificial intelligence according to an embodiment of the invention is shown, where the method includes:
step S1: obtaining a solar pipeline image of a plurality of continuous frames; classifying according to pixel values of pixel points in the pipeline image to obtain abnormal pixel points and normal pixel points; and obtaining a first abnormal region according to the abnormal pixel points.
Cameras are arranged around the solar water heater, so that the cameras can collect front-view images of the solar pipelines, and the integrity of pipeline information in subsequent images is ensured.
Preferably, in order to eliminate the background influence, only pipeline information is reserved, and a plurality of continuous frames of solar water heater images are acquired through a camera. And sending the solar water heater image into a pre-trained pipeline segmentation network to obtain a solar pipeline image. The solar duct image includes a front duct image and a back duct image. Because the effect of defects on the front and back of the pipe is different, independent analysis is required. In the embodiment of the invention, the specific training method of the pipeline segmentation network comprises the following steps:
(1) The image containing the solar pipeline is used as training data. And labeling the solar pipeline pixels as 1, and labeling other pixels as 0 to obtain tag data.
(2) The semantic segmentation network adopts an encoding-decoding structure, and the training data and the label data are input into the network after being normalized. The semantic segmentation encoder is used for extracting the characteristics of the input data and obtaining a characteristic diagram. The semantic segmentation encoder carries out sampling transformation on the feature map and outputs a semantic segmentation result. The mask may be obtained from the semantic segmentation result and the solar pipeline image may be obtained by multiplication with the solar water heater image. It should be noted that, in order to facilitate the subsequent analysis of the solar pipeline image as the combination of all the vacuum pipeline image information, the images of the single pipeline are spliced and stacked into one solar pipeline image, so as to facilitate the subsequent data analysis.
(3) The network is trained using a cross entropy loss function.
In the embodiment of the invention, in order to ensure the quality of the image and avoid the phenomena of specular reflection or diffuse reflection and the like on the surface of the vacuum tube caused by illumination, after the image of the solar water heater is obtained, graying and histogram equalization are carried out for reducing the overeffect of uneven illumination and reflection of the image, and other image preprocessing means can be selected in other embodiments, so that the method is not limited.
The pixel points of the defect area in the image have larger difference with the pixel points of the pipeline, so that the initial classification can be carried out according to the pixel values of the pixel points in the solar pipeline image, and abnormal pixel points and normal pixel points can be obtained. The classification method has various choices, and in the embodiment of the invention, the pixel points are divided into two types by adopting a mean value clustering algorithm, and because the pixel points of the pipeline occupy most, the defective pixel points occupy few, the type with more sample data volume is a normal pixel point, and the type with less sample data volume is an abnormal pixel point. And reserving the abnormal pixel points for connected domain analysis to obtain a plurality of first abnormal areas.
Step S2: obtaining a first gray level co-occurrence matrix of a first abnormal region; obtaining a first texture feature according to the first gray level co-occurrence matrix; and removing the first abnormal region corresponding to the first texture feature similar to the texture feature of the standard pipeline to obtain a second abnormal region.
Because the pollution defect and the crack defect have a larger influence on the solar heat collection efficiency, the chromatic aberration defect is a normal phenomenon in the use process of the pipeline or the inherent characteristic of the pipeline does not influence the solar heat collection efficiency. The first abnormal region is obtained according to the pixel value, so that the first abnormal region simultaneously comprises a dirt defect, a crack defect and a color difference defect, the color difference defect and the dirt defect are regions with inconsistent colors on the pipeline, the first abnormal region and the color difference defect have a mixed recognition phenomenon, and the color difference region in the first abnormal region needs to be removed for accurately analyzing the heat collection efficiency.
It should be noted that, the color difference area is only the color difference effect generated by the pipeline in some local areas and the pipeline itself, so that the texture information of the color difference area should be the same as the texture information of the normal area of the pipeline, and therefore the color difference area can be screened out and removed according to the texture information, which specifically includes:
a first gray level co-occurrence matrix of each first abnormal region is obtained. The gray level co-occurrence matrix may describe texture information of an area, and the method for obtaining the gray level co-occurrence matrix is a conventional technical means, which is not described in detail herein, but only briefly describes the method for obtaining the first gray level co-occurrence matrix in the embodiment of the present invention:
(1) In order to facilitate data analysis, firstly, gray level quantization is performed, and the gray level of a first abnormal region is compressed to a smaller range on the premise of not affecting texture characteristics, in the embodiment of the invention, the gray level of the first abnormal region is compressed to 8 levels so as to reduce the size of a gray level co-occurrence matrix, thereby facilitating subsequent analysis, namely dividing the gray level of each pixel point by 32 and rounding, and converting 0-255 gray levels to 0-7 gray levels.
(2) In the embodiment of the invention, a sliding window with a size of 5*5 is adopted, the sliding step length is 1, and the first abnormal region is processed. For the convenience of analysis, the gray level co-occurrence matrix is calculated in four directions of 0 degree, 45 degrees, 90 degrees and 135 degrees, namely, the data in the four directions are averaged to be used as the data of the final gray level co-occurrence matrix when the sliding window slides once, until the whole first abnormal area is traversed, and the corresponding first gray level co-occurrence matrix is obtained.
The texture features are represented according to various data which can be obtained in the first gray level co-occurrence matrix, and preferably, in order to make the first texture features obtained by the first gray level co-occurrence matrix describe the texture features more clearly, the energy, contrast, local correlation degree and entropy of the first gray level co-occurrence matrix are taken as the first texture features. The energy, contrast, local correlation and entropy are properties of the gray level co-occurrence matrix, and the acquisition method belongs to conventional technical means, and is not repeated herein, but only briefly described in the embodiment of the invention:
(1) Energy: the energy can reflect the thickness degree of the texture, and the specific calculation formula is as follows:
wherein ,in order for the energy to be of a type,in a first gray level co-occurrence matrixElement values at (a).
(2) Contrast ratio: the contrast of the gray level co-occurrence matrix can reflect the local change condition of the first abnormal region, the contrast is the inertia matrix information near the main diagonal of the gray level co-occurrence matrix, the distribution of matrix values is reflected, the depth of textures is reflected, and a specific calculation formula is as follows:
wherein ,for the contrast ratio,in a first gray level co-occurrence matrixElement values at (a).
(3) Local correlation: the local correlation may be analyzed based on the similarity of the pixel gray levels in the row-column direction of the first abnormal region, and the specific calculation formula is as follows:
wherein ,in order to be a local degree of correlation,as the number of gray levels,is the gray-scale average value of the first abnormal region,for the gray variance of the first anomaly region,in a first gray level co-occurrence matrixElement values at (a).
(4) Entropy: the entropy may reflect the texture complexity of the first anomaly region, and the specific calculation formula includes:
wherein ,in order to be an entropy of the water,in a first gray level co-occurrence matrixElement values at (a).
The first texture feature being in vector form, i.e. the firstThe first texture feature of the first abnormal region is that. Comparing the first texture features with standard pipeline textures, and removing similar first abnormal regions to obtain second abnormal regions without color difference regions, wherein the method specifically comprises the following steps:
and obtaining the two-norm difference of the first texture feature and the standard pipeline texture feature, and taking the reciprocal of the two-norm difference as a judging index. When the judging index is larger than a preset judging index threshold value, the corresponding first abnormal region is considered to be similar to the texture characteristics of the normal pipeline, and therefore the first abnormal region belongs to a chromatic aberration region. And removing the first abnormal region corresponding to the judgment index larger than the preset judgment index threshold, namely setting the pixel value of the corresponding pixel point to 0, and obtaining a second abnormal region comprising the dirty region and the crack region.
Step S3: identifying a strip-shaped abnormal region according to the size of the second abnormal region, wherein the strip-shaped abnormal region is taken as a suspected crack region, and other second abnormal regions are dirty regions; obtaining a second gray level co-occurrence matrix of each pixel point of the suspected crack region in a preset neighborhood range, and obtaining a second texture feature set of each suspected crack region according to the second gray level co-occurrence matrix; taking suspected crack areas corresponding to the second texture feature sets with the same similarity as a class of area categories; the region type with the least data amount is used as a crack region, and the other regions are dirty regions.
The second abnormal region comprises a dirty region and a crack region, and the crack region belongs to the strip-shaped region due to the special shape, so that the two regions can be divided for the first time according to the shape, the strip-shaped abnormal region in the second abnormal region is identified, the strip-shaped abnormal region is taken as a suspected crack region, and the other regions are dirty regions. In the embodiment of the present invention, the determination method of the strip-shaped abnormal region is to determine according to the aspect ratio of the smallest circumscribed rectangle of the second abnormal region, the ratio of the longest side to the shortest side of the smallest circumscribed rectangle is taken as the strip-shaped confidence, the second abnormal region larger than the preset strip-shaped confidence threshold is taken as the strip-shaped region, and in other embodiments, other means may be used to identify the strip-shaped region, which is not limited herein.
The suspected crack region includes not only a crack region but also a streak-shaped dirty region due to rain wash or the like. The size and shape of different crack areas are different, but the textures are consistent, the dirty areas are the same, because the dirty areas are all the dirty spots generated by external connection, the dirty areas are different from the texture information of the crack areas, and the distribution quantity of the dirty areas is far greater than that of the crack areas, so that the suspected crack areas can be classified according to the texture information, and the crack areas can be screened according to the data quantity, and the method specifically comprises the following steps:
(1) And obtaining a second gray level co-occurrence matrix in the neighborhood range of each pixel point of the suspected crack region. In the embodiment of the invention, the method for setting the neighborhood range comprises the following steps: and (3) making a vertical line which is perpendicular to the longest side of the minimum circumscribed rectangle of the suspected crack region by passing through the target pixel point, taking the target pixel point as the center on the vertical line, and selecting 7 pixel points on two sides respectively, namely forming a neighborhood range by the selected 7 pixel points and the target pixel point. The size of the second gray level co-occurrence matrix is set to 10×10, i.e., the gray level in the suspected crack region is classified into 10 gray levels. The second gray level co-occurrence matrix in the neighborhood range describes the probability of occurrence of different gray level value combinations in the vertical line direction, and represents the single neighborhood pixel gray level distribution characteristics and distribution characteristics of the target pixel point in the vertical line.
(2) And converting the second gray level co-occurrence matrix into a one-dimensional space to obtain a one-dimensional feature vector. And taking the set of the one-dimensional feature vectors of all the pixel points in the suspected crack area as a second texture feature set.
(3) Because the pipeline is a product with quality overstretched after safety inspection, and a dirty area is easier to form in an outdoor environment, the dirty area distribution is far larger than that of a crack area, so that the suspected crack areas can be classified according to the second texture feature sets, the suspected crack areas corresponding to the second texture feature sets with the same similarity are used as area types, the area types with the least data amount are used as the crack areas, and the other areas are dirty areas. In the embodiment of the invention, any two second texture feature sets are analyzed by adopting a maximum weighted holding difference method to obtain feature differencesObtaining the similarity according to a similarity obtaining formula：. Similarity degreeThe larger the two second texture feature sets are, the more similar the two second texture feature sets are, and in other embodiments, the feature differences may be obtained according to other methods, such as a maximum mean difference method, so as to calculate the similarity, which is not limited herein. In the embodiment of the invention, cluster analysis is performed according to the similarity, and each cluster is a type of region category.
Step S4: obtaining color information of the dirty region, and obtaining the dirty thickness according to the color information; obtaining overall fouling characteristics according to the fouling thickness and the area of the fouling area; obtaining a crack change characteristic according to the size change of the crack region in the continuous multi-frame solar pipeline image and the pixel value change; taking the crack change characteristics and the area of the crack area as integral crack characteristics; and predicting solar heat collection efficiency according to the integral dirt characteristics, the integral crack characteristics and the current environmental data.
According to the processes of step S2 and step S3, a dirty region and a cracked region can be obtained. Because the dirty areas have different effects on the heat collection efficiency at the front and back of the pipeline, different weights can be set for overall analysis. Because the influence of pipeline cracks on heat collection efficiency is mainly reflected in heat loss, the influence of different positions of the front surface and the back surface on the heat collection efficiency is the same.
The color information of the dirty region can effectively indicate the depth of the dirty, namely the darker the dirty region, the thicker the dirty region, and the dirty thickness can be obtained according to the color information, and the method specifically comprises the following steps:
the RGB image of the dirty region is converted into a HIS color space with saturation and brightness information in the HIS color space as color information. The conventional technical means for the HIS color space conversion will not be described in detail herein.
And obtaining the dirt thickness of each dirt area according to a dirt thickness calculation formula. The dirt thickness calculation formula includes:
wherein ,In order for the thickness of the soil to be the same,for the brightness information, the brightness information is used,is saturation information.
And obtaining a front dirty region and a back dirty region according to the front pipeline image and the back pipeline image. And respectively adjusting the area and the dirt thickness of the front dirt area according to the preset front dirt weight. And respectively adjusting the area and the dirt thickness of the back dirt area according to the preset back dirt weight.
Accumulating the area of the front dirt area and the area of the back dirt area after adjustment to obtain overall dirt area information; accumulating and averaging the adjusted dirt thickness of the front dirt area and the dirt thickness of the back dirt area to obtain overall dirt thickness information; the overall dirty area information and the overall dirty thickness information are taken as the overall dirty characteristics. I.e. overall dirty area informationThe method comprises the following steps:
wherein ,as the overall dirty area information,is the firstThe area of the individual front side dirty regions,is the firstThe area of the individual back side soiled areas,as a front-side soil weight,is the backside fouling weight. In an embodiment of the present invention, in the present invention,。
the thickness of the whole dirt isThe method comprises the following steps:
wherein ,for the thickness of the whole dirt, the thickness of the dirt,is the firstThe soil thickness of the individual frontal soil areas,is the firstThe thickness of the soil in the rear soil area,for the number of back side soiled areas,is the number of frontal dirty areas.
I.e. the overall soil is characterized by。
The crack area not only can generate heat loss to the pipeline, but also can be gradually unfolded along with time change, so that the crack is lengthened and deepened, and therefore, dynamic analysis is required to be carried out on the crack, and the influence of the crack is analyzed more accurately. The method specifically comprises the following steps:
and obtaining crack change characteristics according to the size change of the crack region in the continuous multi-frame solar pipeline image and the pixel value size change. It should be noted that, the data acquisition time can be set according to task requirements, and the continuous multi-frame provided by the embodiment of the invention is a data sequence with a plurality of longer acquisition time periods, and in the embodiment of the invention, the sampling time is set to be one week, that is, each frame in the continuous multi-frame solar pipeline image represents the pipeline information of the current week. When the heat collection efficiency of the current week is analyzed, the current frame image and the previous frame image form an image sequence for dynamically analyzing crack changes, and the sequence length can be automatically set according to task requirements and is not described herein.
And obtaining the size change according to a size change formula. The size change formula includes:
wherein ,is the firstThe size of the individual crack areas varies,is the firstArea of the crack region in the frame solar pipe image,for the number of frames of the solar pipeline image, i.eIs the firstArea of the crack region in the frame solar pipe image,is the firstArea of crack region in the frame solar pipe image. The dimensional change represents a dynamic change in crack size.
And obtaining the pixel value size change according to a pixel value size change formula. The pixel value size change formula includes:
wherein ,is the firstThe pixel value of each crack region varies in size,is the firstCracking in frame solar pipeline imagesThe average pixel value of the texture region,is the number of frames of the solar pipeline image. The pixel value size change represents the dynamic change condition of the crack depth.
The product of the size change of the dimension and the size change of the pixel value is taken as the crack change characteristic. Further cumulatively averaging the crack variation characteristics of each crack with the cumulated area of the crack regionThe crack is taken as integral crack characteristics)。
Solar heat collection efficiency can be predicted according to historical data according to the overall dirt characteristics, the overall crack characteristics and the current environmental data. Preferably, the dirty characteristic, the crack characteristic and the current environmental data are input into a pre-trained heat collection efficiency prediction network, and the solar heat collection efficiency is output. The heat collection efficiency prediction network adopts a common deep neural network, and a specific training method is a conventional technical means and is not described in detail herein.
The staff can judge the working condition of current solar water heater according to solar heat collection efficiency, and when the working condition is not ideal, the defect is explained and the influence of heat collection efficiency is great, needs the staff to overhaul solar water heater.
In summary, in the embodiment of the present invention, the first determination is performed according to the pixel value of the image pixel point, so as to obtain the first abnormal region. And eliminating the color difference area according to the first texture characteristics to obtain a second abnormal area. And obtaining a suspected crack region according to the size and morphology of the second abnormal region, classifying according to the second texture features in the suspected crack region, and screening out the crack region and the dirty region. And analyzing the heat collection efficiency of the solar water heater according to the characteristics of the dirty area and the characteristics of the crack area. The embodiment of the invention classifies the defect types and realizes accurate prediction analysis on the heat collection efficiency aiming at the defect types.
The invention also provides a solar heat collection efficiency analysis system based on the artificial intelligence, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, and is characterized in that the processor realizes any one step of the solar heat collection efficiency analysis method based on the artificial intelligence when executing the computer program.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (9)
1. An artificial intelligence-based solar heat collection efficiency analysis method, which is characterized by comprising the following steps:
obtaining a solar pipeline image of a plurality of continuous frames; classifying according to pixel values of pixel points in the solar pipeline image to obtain abnormal pixel points and normal pixel points; obtaining a first abnormal region according to the abnormal pixel points;
obtaining a first gray level co-occurrence matrix of the first abnormal region; obtaining a first texture feature according to the first gray level co-occurrence matrix; removing the first abnormal region corresponding to the first texture feature similar to the standard pipeline texture feature to obtain a second abnormal region;
identifying a strip-shaped abnormal region according to the size of the second abnormal region, wherein the strip-shaped abnormal region is taken as a suspected crack region, and the other second abnormal regions are dirty regions; obtaining a second gray level co-occurrence matrix of each pixel point of the suspected crack region in a preset neighborhood range, and obtaining a second texture feature set of each suspected crack region according to the second gray level co-occurrence matrix; taking the suspected crack areas corresponding to the second texture feature sets with the same similarity as a class of area categories; taking the area category with the minimum data amount as a crack area, and the other areas are the dirty areas;
the method for obtaining the color information of the dirt area comprises the steps of: converting the dirty region into an HIS color space, and taking saturation and brightness information in the HIS color space as the color information;
obtaining the dirt thickness of each dirt area according to a dirt thickness calculation formula; the dirt thickness calculation formula comprises:
wherein ,for the thickness of the soil to be mentioned,for the purpose of the brightness information as described,for the saturationDegree information; obtaining an overall soil characteristic according to the soil thickness and the area of the soil region; obtaining crack change characteristics according to the size change and the pixel value change of the crack region in the continuous multi-frame solar pipeline image; taking the crack change characteristics and the area of the crack area as integral crack characteristics; and predicting solar heat collection efficiency according to the integral dirt characteristics, the integral crack characteristics and the current environmental data.
2. The method for analyzing solar heat collection efficiency based on artificial intelligence according to claim 1, wherein the step of obtaining the solar pipeline images of the continuous multiframe comprises the steps of:
collecting images of the solar water heater of continuous multiframes; sending the solar water heater image into a pre-trained pipeline segmentation network to obtain the solar pipeline image; the solar duct image includes a front duct image and a back duct image.
3. The method of claim 1, wherein the obtaining the first texture feature according to the first gray level co-occurrence matrix comprises:
and taking the energy, contrast, local relativity and entropy of the first gray level co-occurrence matrix as the first texture features.
4. The method of claim 1, wherein said removing the first anomaly region corresponding to the first texture feature similar to the texture feature of the standard pipeline comprises:
obtaining a second norm difference between the first texture feature and the standard pipeline texture feature, and taking the inverse of the second norm difference as a judgment index; and removing the first abnormal region corresponding to the judgment index larger than a preset judgment index threshold value.
5. The artificial intelligence based solar thermal efficiency analysis method of claim 1, wherein the obtaining a second texture feature set for each of the suspected crack regions from the second gray level co-occurrence matrix comprises:
converting the second gray level co-occurrence matrix into a one-dimensional space to obtain a one-dimensional feature vector; and taking the set of the one-dimensional feature vectors of all pixel points in the suspected crack area as the second texture feature set.
6. The method of claim 2, wherein the obtaining the fouling characteristics from the fouling thickness and the area of the fouling area comprises:
acquiring a front dirty region and a back dirty region according to the front pipeline image and the back pipeline image; the area of the front dirty area and the dirty thickness are respectively adjusted according to a preset front dirty weight; respectively adjusting the area of the back dirt area and the dirt thickness according to preset back dirt weight;
accumulating the adjusted area of the front dirt area and the area of the back dirt area to obtain overall dirt area information; accumulating and averaging the adjusted dirt thickness of the front dirt area and the dirt thickness of the back dirt area to obtain overall dirt thickness information; and taking the whole dirt area information and the whole dirt thickness information as the whole dirt characteristics.
7. The artificial intelligence based solar thermal collection efficiency analysis method of claim 1, wherein the obtaining crack variation features from the dimensional and pixel value variations of the crack region in successive frames of the solar pipe image comprises:
obtaining the size change according to a size change formula; the size change formula comprises:
wherein ,is the firstThe size of the dimensions of each of the crack regions varies,is the firstThe area of the crack region in the solar pipe image is framed,a frame number for the solar pipeline image;
obtaining the pixel value change according to a pixel value change formula; the pixel value size variation formula includes:
wherein ,is the firstThe pixel values of each of the crack regions vary in size,is the firstThe average pixel value of the crack region in the solar pipeline image is framed,a frame number for the solar pipeline image;
the crack change feature is a product of the dimensional change and the pixel value change.
8. The artificial intelligence based solar thermal efficiency analysis method of claim 1, wherein the predicting solar thermal efficiency based on the fouling characteristics, the cracking characteristics, and current environmental data comprises:
and inputting the dirt characteristics, the crack characteristics and the current environment data into a pre-trained heat collection efficiency prediction network, and outputting the solar heat collection efficiency.
9. An artificial intelligence based solar thermal efficiency analysis system comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor performs the steps of the method according to any one of claims 1 to 8 when the computer program is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211002558.4A CN115082744B (en) | 2022-08-22 | 2022-08-22 | Artificial intelligence-based solar heat collection efficiency analysis method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211002558.4A CN115082744B (en) | 2022-08-22 | 2022-08-22 | Artificial intelligence-based solar heat collection efficiency analysis method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115082744A CN115082744A (en) | 2022-09-20 |
| CN115082744B true CN115082744B (en) | 2023-09-08 |
Family
ID=83244529
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211002558.4A Active CN115082744B (en) | 2022-08-22 | 2022-08-22 | Artificial intelligence-based solar heat collection efficiency analysis method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115082744B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115858630B (en) * | 2023-02-21 | 2023-05-09 | 新风光电子科技股份有限公司 | Abnormality detection method for energy storage data of energy storage power station |
| CN116664584B (en) * | 2023-08-02 | 2023-11-28 | 东莞市旺佳五金制品有限公司 | Intelligent feedback regulating system for production of thin-wall zinc alloy die casting die |
| CN117437232B (en) * | 2023-12-21 | 2024-03-26 | 山东心传矿山机电设备有限公司 | Water pump impeller appearance defect detection method and system based on image processing |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114723701A (en) * | 2022-03-31 | 2022-07-08 | 南通博莹机械铸造有限公司 | Gear defect detection method and system based on computer vision |
| CN114897898A (en) * | 2022-07-13 | 2022-08-12 | 江苏绿泉装饰工程有限公司 | Wood board quality classification method based on image processing |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN203968037U (en) * | 2014-05-26 | 2014-11-26 | 孙帮聪 | Big-power solar collecting electricity-generating apparatus |
| CN112588764A (en) * | 2020-11-28 | 2021-04-02 | 李振环 | Multi-robot collaborative cleaning method and device on photovoltaic module based on artificial intelligence |
| CN114897918B (en) * | 2022-07-13 | 2022-09-27 | 南通同欧智能装备科技有限公司 | Artificial intelligence-based photovoltaic cleaning robot brush power adjustment method and system |
-
2022
- 2022-08-22 CN CN202211002558.4A patent/CN115082744B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114723701A (en) * | 2022-03-31 | 2022-07-08 | 南通博莹机械铸造有限公司 | Gear defect detection method and system based on computer vision |
| CN114897898A (en) * | 2022-07-13 | 2022-08-12 | 江苏绿泉装饰工程有限公司 | Wood board quality classification method based on image processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115082744A (en) | 2022-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115082744B (en) | Artificial intelligence-based solar heat collection efficiency analysis method and system | |
| CN101996405B (en) | Method and device for rapidly detecting and classifying defects of glass image | |
| CN109376792A (en) | Photovoltaic cell open defect classification method based on multichannel residual error neural network | |
| CN105844621A (en) | Method for detecting quality of printed matter | |
| CN110146791A (en) | A kind of corona detection method based on image procossing | |
| CN101059425A (en) | Method and device for identifying different variety green tea based on multiple spectrum image texture analysis | |
| CN112588764A (en) | Multi-robot collaborative cleaning method and device on photovoltaic module based on artificial intelligence | |
| CN114973032B (en) | Deep convolutional neural network-based photovoltaic panel hot spot detection method and device | |
| Ni et al. | An improved Faster R-CNN for defect recognition of key components of transmission line | |
| CN116703911B (en) | LED lamp production quality detecting system | |
| CN111178177A (en) | Cucumber disease identification method based on convolutional neural network | |
| CN109359554A (en) | A kind of forest fire recognition methods based on more composograph processing techniques | |
| CN115170475A (en) | Non-woven fabric defect detection method based on deep learning method | |
| CN113313107B (en) | Intelligent detection and identification method for multiple types of diseases on cable surface of cable-stayed bridge | |
| CN116805302A (en) | Cable surface defect detection device and method | |
| CN114677336B (en) | Curtain wall panel damage identification method based on infrared image | |
| CN116664565A (en) | Hidden crack detection method and system for photovoltaic solar cell | |
| CN116485785A (en) | Surface defect detection method for solar cell | |
| CN109598681A (en) | The reference-free quality evaluation method of image after a kind of symmetrical Tangka repairs | |
| CN113724842A (en) | Cervical tissue pathology auxiliary diagnosis method based on attention mechanism | |
| CN114565607A (en) | Fabric defect image segmentation method based on neural network | |
| CN113313678A (en) | Automatic sperm morphology analysis method based on multi-scale feature fusion | |
| CN110766675B (en) | Solar cell panel defect detection method | |
| CN201041547Y (en) | Device for analyzing and recognizing different types of green teas based on multi-spectrum image texture | |
| CN117218101A (en) | Composite wind power blade defect detection method based on semantic segmentation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230815 Address after: Building 217, W2-B, Gaoxin Industrial Village, No. 025 Gaoxin South Fourth Road, Gaoxin Community, Yuehai Street, Nanshan District, Shenzhen City, Guangdong Province, 518000 Applicant after: Shenzhen Shuosheng Digital Energy Co.,Ltd. Address before: 226000 Hui Feng Zhen He Qun Cun, Qidong City, Nantong City, Jiangsu Province Applicant before: QIDONG LUOYUAN PV EQUIPMENT Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |