CN114965315A - Rock mass damage degradation rapid evaluation method based on hyperspectral imaging - Google Patents
Rock mass damage degradation rapid evaluation method based on hyperspectral imaging Download PDFInfo
- Publication number
- CN114965315A CN114965315A CN202210538808.XA CN202210538808A CN114965315A CN 114965315 A CN114965315 A CN 114965315A CN 202210538808 A CN202210538808 A CN 202210538808A CN 114965315 A CN114965315 A CN 114965315A
- Authority
- CN
- China
- Prior art keywords
- rock mass
- damage
- hyperspectral
- degradation
- area
- 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.)
- Pending
Links
- 239000011435 rock Substances 0.000 title claims abstract description 139
- 230000015556 catabolic process Effects 0.000 title claims abstract description 84
- 238000006731 degradation reaction Methods 0.000 title claims abstract description 84
- 238000011156 evaluation Methods 0.000 title claims abstract description 77
- 238000000701 chemical imaging Methods 0.000 title claims abstract description 21
- 238000010521 absorption reaction Methods 0.000 claims abstract description 122
- 230000003595 spectral effect Effects 0.000 claims abstract description 89
- 238000013210 evaluation model Methods 0.000 claims abstract description 47
- 238000000034 method Methods 0.000 claims abstract description 29
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 15
- 238000007637 random forest analysis Methods 0.000 claims abstract description 12
- 238000006253 efflorescence Methods 0.000 claims description 16
- 230000006866 deterioration Effects 0.000 claims description 15
- 238000002310 reflectometry Methods 0.000 claims description 12
- 231100000809 Damage Assessment Model Toxicity 0.000 claims description 11
- 238000005185 salting out Methods 0.000 claims description 9
- 230000008030 elimination Effects 0.000 claims description 8
- 238000003379 elimination reaction Methods 0.000 claims description 8
- 241000282461 Canis lupus Species 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 5
- 238000012360 testing method Methods 0.000 claims description 4
- 238000012795 verification Methods 0.000 claims description 3
- 206010037844 rash Diseases 0.000 description 10
- 239000000126 substance Substances 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 238000011065 in-situ storage Methods 0.000 description 3
- 230000002265 prevention Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000004566 building material Substances 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000010298 pulverizing process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/211—Schema design and management
- G06F16/212—Schema design and management with details for data modelling support
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Databases & Information Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- General Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Educational Administration (AREA)
- General Engineering & Computer Science (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Quality & Reliability (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Game Theory and Decision Science (AREA)
- Biomedical Technology (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Software Systems (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides a method for rapidly evaluating rock mass damage degradation based on hyperspectral imaging, which comprises the following steps: collecting hyperspectral information of an area to be evaluated, and acquiring and storing hyperspectral data; selecting a specific area, corresponding hyperspectral data and a damage degradation grade from the evaluated area, and adding a damage degradation label to the hyperspectral data of the rock mass in the specific area according to the damage degradation grade to obtain an original evaluation database; extracting spectral absorption characteristic parameters of hyperspectral data of a specific area, and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels; constructing a rock mass damage evaluation model according to a random forest algorithm, and training and optimizing the rock mass damage evaluation model by adopting a new sample evaluation database; and importing the hyperspectral data of the area to be evaluated into the optimized rock mass damage evaluation model, and acquiring the rock mass damage degradation grade of the area to be evaluated. The method realizes the pixel-level rapid evaluation of the large-area rock mass damage degradation, and has high evaluation precision.
Description
Technical Field
The invention relates to the technical field of rock mass damage evaluation, in particular to a method for quickly evaluating rock mass damage degradation based on hyperspectral imaging.
Background
The rock mass is an aggregate of rock materials, and the rock mass is exposed to the natural environment for a long time, so that the rock mass is influenced by physical, chemical, biological and other factors to generate weathering, and further the service life of the building material is influenced. Therefore, damage and deterioration evaluation of rock mass has been an important issue in geological disaster research.
However, at present, most artificial local measurement methods are used for damage and degradation evaluation of rock mass, and the methods have the problems of low efficiency, susceptibility to subjective factors of evaluation results and the like. Therefore, a method for rapidly and accurately evaluating damage and deterioration of a rock body is needed.
Disclosure of Invention
In view of the above, it is necessary to provide a method for rapidly evaluating damage and deterioration of a rock mass based on hyperspectral imaging.
A method for rapidly evaluating rock mass damage degradation based on hyperspectral imaging comprises the following steps: collecting hyperspectral information of an area to be evaluated, acquiring hyperspectral data and storing the hyperspectral data; selecting a specific area from the evaluated areas, acquiring hyperspectral data and damage degradation grade of rock mass of the specific area, adding a damage degradation label to the hyperspectral data of the rock mass of the specific area according to the damage degradation grade, and acquiring an original evaluation database; extracting spectral absorption characteristic parameters of hyperspectral data of the rock mass in the specific area, and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels; constructing a rock mass damage evaluation model according to a random forest algorithm, and training and optimizing the rock mass damage evaluation model by adopting the new sample evaluation database; and importing the hyperspectral data of the area to be evaluated into the optimized rock mass damage evaluation model, and acquiring the rock mass damage degradation grade of the area to be evaluated.
Further, collecting hyperspectral information of all areas to be evaluated by adopting a portable staring hyperspectral imager; during collection, a fixed parameter collection mode is selected, the collected wave band is 980-1700 nm, the interval channel is 5nm, and the total number of wave bands is 145.
Further, when the portable staring type hyperspectral imager is used for collecting hyperspectral information, the object distance is kept consistent, the error is not more than 1m, and the object distance is the distance between an eyepiece of the portable staring type hyperspectral imager and a rock mass of an area to be evaluated.
Further, the specific region is a region having no damage degradation characteristics and a region having significant damage degradation characteristics.
Further, the damage deterioration grades are three grades of heavy, medium and light, respectively corresponding to chalking, salting-out weathering, chemical weathering and biological weathering.
Further, when the original evaluation database is obtained, the waveband range of the hyperspectral data constructed by the original evaluation database is 1000-1680 nm, and the total number of the hyperspectral data is 137.
Further, the extracting spectral absorption characteristic parameters of the hyperspectral data of the rock mass in the specific area, and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels specifically include: performing envelope elimination on hyperspectral data in the original evaluation database by using a Python algorithm, highlighting spectral absorption characteristic parameters of the rock mass, and extracting the highlighted spectral absorption characteristic parameters; and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels.
Further, the spectral absorption characteristic parameters comprise spectral reflectivity, spectral absorption valley position, spectral absorption depth, spectral absorption width, absorption valley area, absorption symmetry and spectral absorption index; wherein, the spectral absorption valley position is the wavelength at the lowest position of the reflectivity in the absorption valley after envelope removal, and is defined as:
AP＝λ(ρ min )；
in the formula, rho is the reflectivity after the envelope curve is removed; the spectral absorption depth reflects the degree of absorption, defined as:
AD＝1-ρ AP ；
the spectral absorption width is a spectral depth at half of the maximum absorption depth; the absorption valley area is the area of an area enclosed by the absorption valley curve at the base line, and is defined as:
the absorption symmetry is that the area S of the right side region is bounded by the perpendicular line passing through the absorption valley position R And left region S L The common logarithm of the ratio, defined as:
the spectral absorption index is the absorption valley position lambda between absorption valley regions of the original spectral curve M Is defined as the inverse of the ratio of reflectance to baseline value of:
and the base line value is a value corresponding to the absorption valley position on a connecting line of two end points of the interval curve.
Further, a rock mass damage evaluation model is constructed according to a random forest algorithm, the new sample evaluation database is adopted, and the rock mass damage evaluation model is trained and optimized, and the method specifically comprises the following steps: constructing a rock mass damage evaluation model according to a random forest algorithm; and taking the spectral absorption characteristic parameters in the new sample evaluation database as input, taking the damage degradation label as output, importing the input into a rock mass damage evaluation model, and taking a test set as follows: training the rock mass damage evaluation model according to the proportion of 7:3 in the verification set; and adjusting and optimizing the trained rock mass damage evaluation model to obtain the optimized rock mass damage evaluation model.
Further, carrying out over-parameter optimization on the rock mass damage evaluation model by adopting a wolf algorithm, and carrying out precision evaluation on the rock mass damage evaluation model by adopting confusion matrix accuracy and a Kappa coefficient.
Compared with the prior art, the invention has the advantages and beneficial effects that: the method comprises the steps of acquiring hyperspectral information of a region to be evaluated, acquiring and storing hyperspectral data, determining a specific region in the evaluated region, acquiring hyperspectral data and damage degradation grade of the specific region, adding damage degradation labels to the hyperspectral data to form an original evaluation database, extracting spectral absorption characteristic parameters of the hyperspectral data of rock masses of the specific region, constructing a new sample evaluation database based on the spectral absorption characteristic parameters and the corresponding damage degradation labels, constructing a rock mass damage evaluation model according to a random forest algorithm, training and optimizing through the new sample evaluation database, importing the hyperspectral data of the region to be evaluated into the optimized rock mass damage evaluation model to acquire the rock mass damage degradation grade of the region to be evaluated, and performing in-situ, rapid damage evaluation on the rock masses to realize pixel-level rapid evaluation of damage degradation of large-area rock masses, the method has the advantages of saving workload, having high evaluation precision and providing data support for the construction of buildings and structures and the prevention and treatment of geological disasters.
Drawings
FIG. 1 is a schematic flow chart of a method for rapidly evaluating damage and deterioration of a rock mass based on hyperspectral imaging in one embodiment;
FIG. 2 is a block flow diagram of a method for rapidly evaluating damage and deterioration of a rock mass based on hyperspectral imaging in one embodiment;
FIG. 3 is a graphical representation of spectral absorption characteristics in one embodiment.
Detailed Description
In order that the invention may be more clearly understood, the following detailed description of the invention is given with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1, a method for rapidly evaluating damage and deterioration of a rock mass based on hyperspectral imaging is provided, which comprises the following steps:
and S101, collecting hyperspectral information of the area to be evaluated, and acquiring and storing hyperspectral data.
Specifically, a hyperspectral imaging technology is adopted to collect hyperspectral information of an area to be evaluated, hyperspectral data of the area to be evaluated are obtained and stored, and the damage and degradation conditions of the rock mass of the area to be evaluated are evaluated through the hyperspectral data.
The method comprises the steps of collecting hyperspectral information of all areas to be evaluated by a portable staring hyperspectral imager, selecting a fixed parameter collection mode of the portable staring hyperspectral imager during collection, wherein the collected wave bands are 980-1700 nm, the spacing channels are 5nm, and the total number of the wave bands is 145.
Specifically, in order to better collect hyperspectral information of an area to be evaluated, a portable staring hyperspectral imager capable of carrying out remote and large-range target objects can be adopted, when data collection is carried out through the imager, a fixed parameter collection mode of the imager is selected, the consistency of the obtained data is ensured, the collected wave band range is 980-1700 nm, the collected interval channels are 5nm, and 145 wave bands are counted.
When the portable staring type hyperspectral imager is adopted for hyperspectral information acquisition, the object distance is kept consistent, the error is not more than 1m, and the object distance is the distance between an eyepiece of the portable staring type hyperspectral imager and a rock mass of an area to be evaluated.
Specifically, when a portable staring type hyperspectral imager is adopted for hyperspectral information acquisition, in order to ensure the consistency of hyperspectral imaging precision, the consistency of object distances needs to be ensured, wherein the object distances are the distances between eyepieces of the portable staring type hyperspectral imager and rock masses of an area to be evaluated, and the offset error of the object distances is not more than 1 m.
And S102, selecting a specific area from the evaluated areas, acquiring hyperspectral data and damage degradation grade of rock mass of the specific area, adding a damage degradation label to the hyperspectral data of the rock mass of the characteristic area according to the damage degradation grade, and acquiring an original evaluation database.
Specifically, the evaluated area may be a rock mass damage deterioration level determined by expert evaluation and with reference to "engineering rock mass grading standard" (GB 50218-94). And selecting a specific area in the evaluated area, such as an area with obvious damage degradation characteristics or an area without damage degradation. And acquiring hyperspectral data and damage degradation grades of a specific area, and marking corresponding damage degradation labels, such as a pulverization strong grade, a salting-out weathering medium grade, a chemical weathering weak grade and the like, on the hyperspectral data of the rock mass of the specific area according to the damage degradation grades. And constructing an original evaluation database according to the hyperspectral data of the specific area and the corresponding damage degradation label for subsequent training of the model.
Wherein the specific region is a region with no damage degradation characteristics and a region with obvious damage degradation characteristics.
Specifically, the specific region selected from the evaluated regions can be a region with no damage degradation characteristics and a region with obvious damage degradation characteristics, and is selected according to the difference of the evaluated regions, so that a region with more comprehensive rock damage degradation characteristics can be conveniently obtained and used for constructing an original evaluation database.
Wherein, the damage deterioration grades are three grades of strong, medium and weak respectively corresponding to efflorescence, salting-out weathering, chemical weathering and biological weathering.
In particular, deterioration due to damage to rock mass includes physical weathering, which is subdivided into chalking and salting-out weathering, chemical weathering and biological weathering. Therefore, the damage degradation grades of the rock mass comprise three grades of heavy, medium and light, which respectively correspond to chalking, salting-out weathering, chemical weathering and biological weathering, and are respectively as follows: severe chalking, moderate chalking, mild chalking, severe salting-out efflorescence, moderate salting-out efflorescence, mild salting-out efflorescence, severe chemical efflorescence, moderate chemical efflorescence, mild chemical efflorescence, severe biological efflorescence, moderate biological efflorescence and mild biological efflorescence 12 damage degradation grades.
When the original evaluation database is obtained, the wave band range of the hyperspectral data constructed by the original evaluation database is 1000-1680 nm, and the total wave band is 137.
Specifically, the head and tail wave bands of the hyperspectral data are affected by background noise, and the accuracy of the data is insufficient, so that when the original evaluation database is obtained, the head and tail wave bands of the acquired hyperspectral data need to be removed, 137 wave bands with the wave band range of 1000-1680 nm are obtained, the influence of the background noise can be weakened, and the accuracy of the data in the original evaluation database is improved.
And S103, extracting spectral absorption characteristic parameters of the hyperspectral data of the rock mass in the specific area, and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels.
Specifically, according to the selected specific area, spectral absorption characteristic parameters corresponding to hyperspectral data are extracted, a new sample evaluation database is constructed according to the spectral absorption characteristic parameters and corresponding damage degradation labels, the damage degradation condition of rock masses in the specific area can be obtained according to the spectral absorption characteristic parameters, a subsequent model is trained through data of the new sample evaluation database, and the damage degradation condition of the area to be evaluated can be evaluated conveniently through the trained model.
In addition, because the rock mass is greatly influenced by the geographic position and the geological information, the new sample evaluation database is only suitable for the evaluation region and the adjacent region thereof, and the evaluation database needs to be reconstructed for the region with the greatly changed geographic position and the geological information, so that the accuracy and the reliability of the evaluation result are improved.
Wherein, step S103 specifically includes: performing envelope elimination on hyperspectral data in the original evaluation database by using a Python algorithm, highlighting spectral absorption characteristic parameters of the rock mass, and extracting the highlighted spectral absorption characteristic parameters; and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels.
In particular, the removal of the envelope effectively highlights the absorption, reflection and emission characteristics of the spectral curves and normalizes them to a uniform spectral background, with the use of comparisons of characteristic values with other spectral curves. When envelope elimination is performed, elimination of envelope in the hyperspectral data is achieved by adopting a Python algorithm through a code mode, for example: converting the hyperspectral data into a hyperspectral curve, obtaining all maximum value points on the hyperspectral curve through derivation, and comparing the maximum value points to obtain the maximum value point in the maximum value points; taking the maximum point as an end point of the envelope curve, calculating the slope of a connecting line between the point and each maximum value in the wavelength increasing direction, taking the maximum point of the slope as the next end point of the envelope curve, and taking the point as a starting point to circulate until the last point on the hyperspectral curve; taking the maximum value point as an end point of the envelope curve, taking the slope of each minimum value connecting line in the direction of reducing the wavelength, taking the minimum slope point as the next end point of the envelope curve, and taking the point as a starting point to circulate until the starting point on the hyperspectral curve; all the end points are connected along the direction of wavelength increase to form an envelope curve, the reflectivity of the corresponding wave band on the envelope curve is removed by using the actual spectral reflectivity, the value after the envelope elimination method is normalized can be obtained, and the envelope elimination of the high spectral curve is realized.
Specifically, envelope lines of hyperspectral data in an original evaluation database are removed through a Python algorithm, spectral absorption characteristic parameters of rock masses are highlighted, the spectral absorption characteristic parameters are convenient to extract subsequently, a new sample evaluation database is constructed according to the extracted spectral absorption characteristic parameters and damage degradation labels, and subsequent model training is performed through the new sample database, so that a better training effect can be obtained, and the accuracy and the reliability of the model are ensured.
In one embodiment, as shown in FIG. 3, the spectral absorption characteristic parameters include, but are not limited to, spectral reflectance, spectral absorption valley location, spectral absorption depth, spectral absorption width, absorption valley area, absorption symmetry, and spectral absorption index; the spectral absorption valley position is the wavelength at the lowest position of the reflectivity in the absorption valley after the envelope is removed, and is defined as:
AP＝λ(ρ min )；
in the formula, rho is the reflectivity after the envelope curve is removed; the spectral absorption depth reflects the degree of absorption, defined as:
AD＝1-ρ AP ；
the spectral absorption width is the spectral depth at half of the maximum absorption depth; the absorption valley area is the area of the region enclosed by the absorption valley curve at the base line, and is defined as:
absorption symmetry is the area S of the right side region bounded by the perpendicular to the position of the over-absorption valley R And left region S L The common logarithm of the ratio, defined as:
the spectral absorption index is the absorption valley position lambda between absorption valley regions of the original spectral curve M Is defined as the inverse of the ratio of reflectance to baseline value of:
wherein, the base line value is the corresponding value of the absorption valley position on the connecting line of the two end points of the interval curve.
And S104, constructing a rock mass damage evaluation model according to a random forest algorithm, and training and optimizing the rock mass damage evaluation model by adopting a new sample evaluation database.
Specifically, a rock mass damage evaluation model is constructed by adopting a random forest algorithm, and the rock mass damage evaluation model is trained and optimized by utilizing a new sample evaluation database, so that the rock mass damage condition of the area to be evaluated can be determined through the rock mass damage evaluation model.
Wherein, step S104 specifically includes: constructing a rock mass damage evaluation model according to a random forest algorithm; and (3) taking the spectral absorption characteristic parameters in the new sample evaluation database as input, taking the damage degradation label as output, importing the input into a rock mass damage evaluation model, and taking a test set as follows: training the rock mass damage evaluation model according to the verification set ratio of 7: 3; and adjusting and optimizing the trained rock mass damage evaluation model to obtain the optimized rock mass damage evaluation model.
The method comprises the steps of conducting super-parameter optimization on a rock mass damage evaluation model by adopting a wolf algorithm, and conducting precision evaluation on the rock mass damage evaluation model by adopting confusion matrix accuracy and a Kappa coefficient.
Specifically, the gray wolf algorithm has the characteristics of strong convergence performance, few parameters, easiness in implementation and the like, so that the gray wolf algorithm can be used for carrying out over-parameter optimization on the rock mass damage assessment model, adjustment and optimization of the rock mass damage assessment model are realized, meanwhile, accuracy assessment is carried out on the rock mass damage assessment model by adopting confusion matrix accuracy and a Kappa coefficient, the reliability of the rock mass damage assessment model is ensured, and the assessment accuracy is improved.
And S105, importing the hyperspectral data of the area to be evaluated into the optimized rock mass damage evaluation model, and acquiring the rock mass damage degradation grade of the area to be evaluated.
Specifically, the hyperspectral data of the rock mass of the area to be evaluated is imported into an optimized rock mass damage evaluation model, the optimized rock mass damage evaluation model is adopted to evaluate the hyperspectral data of each pixel point of the area to be evaluated, an evaluation result is output, the rock mass damage degradation grade of the area to be evaluated is obtained, and in-situ, lossless and rapid damage evaluation of the rock mass is achieved, so that building or structure construction and prevention and treatment of geological disasters are facilitated. In addition, the damage degradation cloud picture of the rock mass of the area to be evaluated can be drawn according to the evaluation result, so that the overall damage degradation condition in the area to be evaluated can be conveniently and intuitively checked.
In the embodiment, hyperspectral data is acquired and stored by acquiring hyperspectral information of an area to be evaluated, a specific area is determined in the evaluated area, hyperspectral data and damage degradation grade of the specific area are acquired, a damage degradation label is added to the hyperspectral data to form an original evaluation database, the spectral absorption characteristic parameters of the hyperspectral data of rock mass in the specific area are extracted, a new sample evaluation database is constructed based on the spectral absorption characteristic parameters and the corresponding damage degradation label, a rock mass damage evaluation model is constructed according to a random forest algorithm, training and optimization are performed through the new sample evaluation database, the hyperspectral data of the area to be evaluated is led into the optimized rock mass damage evaluation model to acquire the rock mass damage degradation grade of the area to be evaluated, in-situ, lossless and rapid damage evaluation can be performed on the rock mass, and pixel-level rapid evaluation of large-area rock mass damage degradation is realized, the method has the advantages of saving workload, having high evaluation precision and providing data support for the construction of buildings and structures and the prevention and treatment of geological disasters.
In one embodiment, a gaze-type hyperspectral imager is used for collecting hyperspectral information of a certain area, the collection mode is fixed parameter collection, the area of a collected rock mass is about 20 x 20m, the collection object distance is 4.5m, the collection wave band is 980-1700 nm, the spacing channels are 5nm, and the total number of wave bands is 145. The hyperspectral imager has the advantages that the acquisition range of the hyperspectral imager cannot cover the whole area, so that the hyperspectral image of the whole area is acquired by adopting a mode of multi-time acquisition and later-stage splicing.
And (3) performing rock mass damage degradation grade evaluation on a specific small-range region with obvious damage degradation characteristics by an expert, grading the damage degradation grade of the specific region by combining with national standards, and attaching corresponding damage degradation grade labels to hyperspectral data of different damage degradation grade regions.
And (3) cutting the head and tail wave bands of the acquired rock mass hyperspectral data by using a wave band cutting function in Analysis software to weaken the influence of background noise, wherein the finally reserved wave band range is 1000-1680 nm, the spacing channel is 5nm, and the total number of the wave bands is 137.
And (4) taking the hyperspectral data cut from the rock mass region of the evaluated grading region as an original evaluation database.
The method comprises the following steps of preprocessing an original evaluation database, extracting spectral absorption characteristic parameters, and training a rock mass damage evaluation model by using a stochastic forest algorithm, and specifically comprises the following steps:
performing envelope elimination on hyperspectral data in sample data by using a Python algorithm to highlight spectral absorption characteristics of a rock mass and extract spectral absorption characteristic parameters of damage and degradation of the rock mass, wherein the spectral absorption characteristic parameters include but are not limited to spectral reflectivity, absorption valley positions, absorption depths, absorption widths, absorption valley areas, absorption symmetries and spectral absorption indexes, and the spectral absorption characteristic parameters of partial pixel points are shown in the following table:
table 1 spectral absorption characteristic parameters of partial pixel points
| Parameter/sequence number | 1 | 2 | 3 | 4 | 5 | 6 |
| Absorption valley reflectivity | 0.872547 | 0.87631 | 0.81785 | 0.844737 | 0.83121 | 0.793159 |
| Absorption valley position | 1435 | 1425 | 1425 | 1430 | 1420 | 1410 |
| Depth of absorption | 0.127453 | 0.12369 | 0.18215 | 0.155263 | 0.16879 | 0.206841 |
| Absorption width | 0.953576 | 0.952168 | 0.95134 | 0.948526 | 0.945041 | 0.945837 |
| Absorption symmetry | 0.478261 | 0.686567 | 0.788889 | 0.547619 | 0.574468 | 0.690141 |
| Spectral absorption index | 1.073105 | 1.071013 | 1.112631 | 1.095952 | 1.104886 | 1.132728 |
| Absorption of grain area | 7.08089 | 6.68836 | 18.11953 | 10.28382 | 9.74465 | 14.48361 |
Adding a damage degradation label to the extracted spectral absorption characteristic parameters to serve as a new sample evaluation database;
taking spectral absorption characteristic parameters as input, taking a damage degradation label as output, importing data in a new sample evaluation database into a rock damage evaluation model, taking 70% of data in the new sample evaluation database as a training set and 30% of data as a test set, and training the rock damage evaluation model;
and optimizing the over-parameters of the trained rock damage assessment model by combining a wolf algorithm, and optimizing the rock damage assessment model by adopting the confusion matrix accuracy and the Kappa coefficient as the evaluation indexes of the model accuracy.
In this embodiment, an envelope curve of the hyperspectral data of the rock mass in the specific region is removed by adopting a Python algorithm, corresponding spectral absorption characteristic parameters are highlighted and extracted, the extracted spectral absorption characteristic parameters are input into a rock mass damage assessment model, the rock mass damage assessment model is trained and optimized, the accuracy of the optimized rock mass damage assessment model is 97.76%, the Kappa coefficient is 0.973, and the model trained by constructing a database by using the hyperspectral data of the specific region as a sample has a good assessment effect.
The present invention is described in further detail with reference to specific embodiments, and the specific embodiments are not to be considered as limited to the description. For those skilled in the art to which the invention pertains, numerous simple deductions or substitutions may be made without departing from the spirit of the invention, which shall be deemed to belong to the scope of the invention.
Claims (10)
1. A method for rapidly evaluating rock mass damage degradation based on hyperspectral imaging is characterized by comprising the following steps:
collecting hyperspectral information of an area to be evaluated, acquiring hyperspectral data and storing the hyperspectral data;
selecting a specific area from the evaluated areas, acquiring hyperspectral data and damage degradation grade of rock mass of the specific area, adding a damage degradation label to the hyperspectral data of the rock mass of the specific area according to the damage degradation grade, and acquiring an original evaluation database;
extracting spectral absorption characteristic parameters of hyperspectral data of the rock mass in the specific area, and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels;
constructing a rock mass damage evaluation model according to a random forest algorithm, and training and optimizing the rock mass damage evaluation model by adopting the new sample evaluation database;
and importing the hyperspectral data of the area to be evaluated into the optimized rock mass damage evaluation model, and acquiring the rock mass damage degradation grade of the area to be evaluated.
2. The method for rapidly evaluating the damage and the degradation of the rock mass based on the hyperspectral imaging is characterized in that a portable staring hyperspectral imager is adopted to collect hyperspectral information of all areas to be evaluated; during collection, a fixed parameter collection mode is selected, the collected wave band is 980-1700 nm, the interval channel is 5nm, and the total number of wave bands is 145.
3. The method for rapidly evaluating rock mass damage and degradation based on hyperspectral imaging according to claim 2 is characterized in that when a portable staring hyperspectral imager is adopted for hyperspectral information acquisition, the object distance is kept consistent, the error is not more than 1m, and the object distance is the distance between an eyepiece of the portable staring hyperspectral imager and a rock mass in an area to be evaluated.
4. The method for rapidly evaluating damage and deterioration of a rock mass based on hyperspectral imaging according to claim 1 is characterized in that the specific area is an area with no damage and deterioration characteristics and an area with obvious damage and deterioration characteristics.
5. The method for rapidly evaluating the damage and the deterioration of the rock mass based on the hyperspectral imaging is characterized in that the damage and deterioration grades are three grades of heavy, medium and light, which respectively correspond to chalking, salting-out weathering, chemical weathering and biological weathering.
6. The method for rapidly evaluating the damage and the degradation of the rock mass based on the hyperspectral imaging as claimed in claim 2 is characterized in that when the original evaluation database is obtained, the waveband range of the hyperspectral data used for constructing the original evaluation database is 1000-1680 nm, and 137 wavebands are counted.
7. The method for rapidly evaluating rock mass damage degradation based on hyperspectral imaging according to claim 1 is characterized in that the steps of extracting the spectral absorption characteristic parameters of the hyperspectral data of the rock mass in the specific area and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels specifically comprise:
performing envelope elimination on hyperspectral data in the original evaluation database by using a Python algorithm, highlighting spectral absorption characteristic parameters of the rock mass, and extracting the highlighted spectral absorption characteristic parameters;
and constructing a new sample evaluation database according to the spectral absorption characteristic parameters and the damage degradation labels.
8. The method for rapidly evaluating rock mass damage degradation based on hyperspectral imaging according to claim 7 is characterized in that the spectral absorption characteristic parameters comprise spectral reflectivity, spectral absorption valley position, spectral absorption depth, spectral absorption width, absorption valley area, absorption symmetry and spectral absorption index;
wherein, the spectral absorption valley position is the wavelength at the lowest position of the reflectivity in the absorption valley after envelope removal, and is defined as:
AP＝λ(ρ min )；
in the formula, rho is the reflectivity after the envelope curve is removed;
the spectral absorption depth reflects the degree of absorption, defined as:
AD＝1-ρ AP ；
the spectral absorption width is a spectral depth at half of the maximum absorption depth;
the absorption valley area is the area of an area enclosed by the absorption valley curve at the base line, and is defined as:
the absorption symmetry is that the area S of the right side region is bounded by the perpendicular line passing through the absorption valley position R And left region S L The common logarithm of the ratio, defined as:
the spectral absorption index is the absorption valley position lambda between absorption valley regions of the original spectral curve M Is defined as the inverse of the ratio of reflectance to baseline value of:
and the base line value is a value corresponding to the absorption valley position on a connecting line of two end points of the interval curve.
9. The hyperspectral imaging-based rock mass damage degradation rapid assessment method according to claim 1 is characterized in that a rock mass damage assessment model is constructed according to a random forest algorithm, the new sample assessment database is adopted to train and optimize the rock mass damage assessment model, and the method specifically comprises the following steps:
constructing a rock mass damage evaluation model according to a random forest algorithm;
and taking the spectral absorption characteristic parameters in the new sample evaluation database as input, taking the damage degradation label as output, importing the input into a rock mass damage evaluation model, and taking a test set as follows: training the rock mass damage evaluation model according to the proportion of 7:3 in the verification set;
and adjusting and optimizing the trained rock mass damage evaluation model to obtain the optimized rock mass damage evaluation model.
10. The method for rapidly evaluating the damage and the degradation of the rock mass based on the hyperspectral imaging as claimed in claim 9 is characterized in that the grey wolf algorithm is adopted to carry out the hyperparametric optimization on the rock mass damage evaluation model, and the confusion matrix accuracy and the Kappa coefficient are adopted to carry out the precision evaluation on the rock mass damage evaluation model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210538808.XA CN114965315A (en) | 2022-05-18 | 2022-05-18 | Rock mass damage degradation rapid evaluation method based on hyperspectral imaging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210538808.XA CN114965315A (en) | 2022-05-18 | 2022-05-18 | Rock mass damage degradation rapid evaluation method based on hyperspectral imaging |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114965315A true CN114965315A (en) | 2022-08-30 |
Family
ID=82984256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210538808.XA Pending CN114965315A (en) | 2022-05-18 | 2022-05-18 | Rock mass damage degradation rapid evaluation method based on hyperspectral imaging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114965315A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117493856A (en) * | 2024-01-03 | 2024-02-02 | 佛山科学技术学院 | Random forest-based method and equipment for analyzing characteristic factors of fruit picking |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20180084479A (en) * | 2017-01-17 | 2018-07-25 | 서울시립대학교 산학협력단 | Apparatus and method for discriminating a concrete status using a hyperspectral image |
| CN109283148A (en) * | 2018-09-30 | 2019-01-29 | 核工业北京地质研究院 | A method of based on spectral information automatic identification rock forming mineral |
| CN110188716A (en) * | 2019-06-04 | 2019-08-30 | 中国煤炭地质总局勘查研究总院 | A kind of mineralization information extracting method and system |
| CN111261236A (en) * | 2020-01-21 | 2020-06-09 | 山东大学 | Tunnel igneous rock weathering degree determining system and method |
| CN113685188A (en) * | 2021-08-16 | 2021-11-23 | 中铁十八局集团有限公司 | TBM tunneling optimization method based on physical characteristics of rock slag |
-
2022
- 2022-05-18 CN CN202210538808.XA patent/CN114965315A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20180084479A (en) * | 2017-01-17 | 2018-07-25 | 서울시립대학교 산학협력단 | Apparatus and method for discriminating a concrete status using a hyperspectral image |
| CN109283148A (en) * | 2018-09-30 | 2019-01-29 | 核工业北京地质研究院 | A method of based on spectral information automatic identification rock forming mineral |
| CN110188716A (en) * | 2019-06-04 | 2019-08-30 | 中国煤炭地质总局勘查研究总院 | A kind of mineralization information extracting method and system |
| CN111261236A (en) * | 2020-01-21 | 2020-06-09 | 山东大学 | Tunnel igneous rock weathering degree determining system and method |
| CN113685188A (en) * | 2021-08-16 | 2021-11-23 | 中铁十八局集团有限公司 | TBM tunneling optimization method based on physical characteristics of rock slag |
Non-Patent Citations (3)
| Title |
|---|
| LIU, CJ 等: "Effects of freeze-thaw treatment on the dynamic tensile strength of granite using the Brazilian test", COLD REGIONS SCIENCE AND TECHNOLOGY, vol. 155, 1 November 2018 (2018-11-01), pages 372 - 332 * |
| 孙美君;柴勃隆;张冬;王征;孙济洲;: "基于近红外高光谱技术的敦煌莫高窟壁画起甲病害风险评估方法", 文物保护与考古科学, no. 04, 15 November 2016 (2016-11-15), pages 1 - 8 * |
| 张立福 等: "《高光谱遥感信息处理》", vol. 1, 30 June 2021, 湖北科学技术出版社, pages: 147 - 148 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117493856A (en) * | 2024-01-03 | 2024-02-02 | 佛山科学技术学院 | Random forest-based method and equipment for analyzing characteristic factors of fruit picking |
| CN117493856B (en) * | 2024-01-03 | 2024-06-14 | 佛山科学技术学院 | Random forest-based method and equipment for analyzing characteristic factors of fruit picking |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108956505B (en) | Method and device for detecting concentration of chlorophyll a in small water body based on Sentinel-2 image | |
| CN108414455B (en) | Crop hail disaster remote sensing monitoring method for agricultural insurance claim settlement | |
| WO2021012898A1 (en) | Artificial intelligence-based agricultural insurance surveying method and related device | |
| CN107895136B (en) | Coal mine area identification method and system | |
| CN111161362A (en) | Tea tree growth state spectral image identification method | |
| CN113420412B (en) | Soil organic carbon content continuous depth distribution extraction method based on imaging spectrum | |
| CN114965315A (en) | Rock mass damage degradation rapid evaluation method based on hyperspectral imaging | |
| Gara et al. | Predicting forest carbon stocks from high resolution satellite data in dry forests of Zimbabwe: exploring the effect of the red-edge band in forest carbon stocks estimation | |
| US20050111692A1 (en) | Cloud cover assessment: VNIR-SWIR | |
| CN111582689A (en) | Evaluation method and system for ecological protection red line protection effect | |
| CN115128013A (en) | Soil organic matter content space prediction evaluation method based on partition algorithm | |
| CN111474117B (en) | Method and device for monitoring crop diseases | |
| CN116645603A (en) | Soybean planting area identification and area measurement method | |
| CN115294147A (en) | Method for estimating aboveground biomass of single trees and forests based on unmanned aerial vehicle laser radar | |
| CN114676636A (en) | Grassland area soil moisture rapid inversion method integrating vegetation and habitat characteristics | |
| CN114397277A (en) | Unmanned aerial vehicle water chlorophyll remote sensing detection system | |
| CN113887493B (en) | Black and odorous water body remote sensing image identification method based on ID3 algorithm | |
| CN110849821A (en) | Black and odorous water body remote sensing identification method based on Bayesian theorem | |
| CN117035174A (en) | Method and system for estimating biomass on single-woodland of casuarina equisetifolia | |
| CN117491293A (en) | High-steep bank slope carbonate rock corrosion rapid evaluation method based on hyperspectrum | |
| Schillaci et al. | Assessment of hail damages in maize using remote sensing and comparison with an insurance assessment: A case study in Lombardy | |
| US20240044783A1 (en) | Generating high-resolution concentration maps for atmospheric gases using geography-informed machine learning | |
| CN114755189B (en) | Feature-optimized self-attention-based hyperspectral satellite LAI inversion method | |
| Deo | Modelling and mapping of above-ground biomass and carbon sequestration in the cool temperate forest of north-east China | |
| CN116596198B (en) | Urban green land biomass calculation method, device, electronic equipment and storage medium |
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 |