CN114120147A - Ecological environment evaluation system for railway tunnel water inrush flow path along line region - Google Patents

Ecological environment evaluation system for railway tunnel water inrush flow path along line region Download PDF

Info

Publication number
CN114120147A
CN114120147A CN202111445915.XA CN202111445915A CN114120147A CN 114120147 A CN114120147 A CN 114120147A CN 202111445915 A CN202111445915 A CN 202111445915A CN 114120147 A CN114120147 A CN 114120147A
Authority
CN
China
Prior art keywords
tunnel
layer
remote sensing
construction
convolution
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.)
Granted
Application number
CN202111445915.XA
Other languages
Chinese (zh)
Other versions
CN114120147B (en
Inventor
曾赛星
杨旭
刘威
张静晓
李玉龙
高鑫
陈宏权
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Jiaotong University
Harbin Normal University
Original Assignee
Shanghai Jiaotong University
Harbin Normal University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Jiaotong University, Harbin Normal University filed Critical Shanghai Jiaotong University
Priority to CN202111445915.XA priority Critical patent/CN114120147B/en
Publication of CN114120147A publication Critical patent/CN114120147A/en
Application granted granted Critical
Publication of CN114120147B publication Critical patent/CN114120147B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Image Processing (AREA)

Abstract

The invention discloses an ecological environment evaluation system for a region along a railway tunnel water inrush flow path, and relates to an ecological environment evaluation system for a region along a railway tunnel water inrush flow path. The invention aims to solve the problems that the influence of the water burst of the existing tunnel on the ecological environment is difficult to evaluate and the rapid dynamic evaluation cannot be realized. The system comprises: the water inrush flow path and buffer area building module is used for determining a flowing path of the tunnel water inrush current and building a water inrush flow path buffer area by taking the flowing path of the tunnel entrance water inrush current as a central line; the multispectral remote sensing image acquisition and preprocessing module is used for acquiring multi-period real-time remote sensing image monitoring data of a tunnel buffer area to be evaluated; the tunnel water burst ecological evaluation index construction and ecological factor index inversion module is used for obtaining remote sensing ecological factor index parameters; the remote sensing ecological improvement model evaluation module for the tunnel water inrush flow path region is used for obtaining the risk condition of the ecological environment of the tunnel water inrush flow path region along the line. The method is used for the field of ecological environment evaluation of the area along the water flowing path of the railway tunnel.

Description

Ecological environment evaluation system for railway tunnel water inrush flow path along line region
Technical Field
The invention relates to an ecological environment evaluation system for a railway tunnel water inrush flow path along a line region.
Background
The tunnel is used as an important structure in railway engineering construction, and plays an increasingly important role in the aspects of improving the line shape of a route, shortening mileage and driving time, improving operation benefits and the like. As a common large-scale auxiliary project in mountain railway construction, a railway tunnel has the engineering characteristics of complex technology, high construction difficulty and the like, and relates to natural environments such as terrain, landform, meteorology, hydrology, ecology, stratum, lithology, geological structure and the like and various natural factors such as engineering geology and the like. The implementation of a large amount of underground engineering can generate lasting influence on the surrounding environment, wherein the environmental problems caused by tunnel water inrush are particularly sudden, according to incomplete statistics, nearly 80% of tunnels such as traffic and the like have more or less sudden water inrush disasters in the construction period or the operation period, the water inrush amount of the tunnels is increased to lower the water level of underground water in the region, so that mountainous areas or basin wetlands are shrunk or disappear, pollutants contained in the water inrush can influence the growth of surface vegetation more, the ecological environment is degraded, the landscape pattern of the region is changed, the influence on the ecological environment around the tunnel area inevitably can be generated, a plurality of ecological environment problems related to the ecological environment are caused, and the living environment of people is damaged.
At present, related researches on railway tunnel construction are less, existing researches mainly focus on the overall influence of the whole railway tunnel construction on the environment and the ground settlement research caused by tunnel water burst, systematic researches on the ecological environment influence of the flowing path area of tunnel water burst water flow are not carried out, and the influence of the water burst on the ecological environment along the flowing path is great.
Therefore, a set of flexible, open and high-applicability technical method is needed to be established to effectively evaluate the ecological environment of the area along the water inrush flow path of the tunnel, support the informatization expression of the ecological environment change in the construction process of the tunnel in the whole life cycle, and solve the problems that the influence of the water inrush of the existing tunnel on the ecological environment is difficult to evaluate and the quick and accurate real-time dynamic evaluation cannot be realized on the basis of checking and verifying the applicability after field investigation.
Disclosure of Invention
The invention aims to solve the problems that the influence of the current tunnel water burst on the ecological environment is difficult to evaluate and the rapid and accurate real-time dynamic evaluation cannot be realized, and provides an ecological environment evaluation system for the area along the water burst flow path of the railway tunnel.
The ecological environment evaluation system for the area along the water inrush flow path of the railway tunnel comprises:
the system comprises a water inrush flow path and buffer area construction module, a multispectral remote sensing image acquisition and preprocessing module, a tunnel water inrush ecological evaluation index construction and ecological index inversion module and a tunnel water inrush flow path region remote sensing ecological improvement model evaluation module;
the water inrush flow path and buffer zone building module is used for determining a flowing path of the water inrush current at the tunnel portal and building a buffer zone along the water inrush flow path by taking the flowing path of the water inrush current at the tunnel portal as a central line;
the multispectral remote sensing image acquisition and preprocessing module is used for acquiring image monitoring data of a tunnel to be analyzed from a remote sensing satellite, and preprocessing the remote sensing image monitoring data to obtain preprocessed multi-period real-time remote sensing image monitoring data of a buffer area along a water inrush flow path of the tunnel to be analyzed;
the tunnel water burst ecological evaluation index construction and ecological factor index inversion module is used for inverting and extracting remote sensing ecological factor index parameters required by evaluation according to preprocessed multi-period real-time remote sensing image monitoring data of a tunnel buffer area to be analyzed;
the remote sensing ecological improvement model evaluation module of the tunnel water inrush flow path line region is used for constructing an optimal remote sensing ecological improvement evaluation model RSEI according to remote sensing ecological factor index parameters required by evaluation extracted by inversion, and obtaining the change conditions of the ecological environment of the tunnel water inrush flow path line region before tunnel construction, during construction and after construction based on the optimal remote sensing ecological improvement evaluation model RSEI.
The invention has the beneficial effects that:
based on a space-attribute-time sequence multi-dimensional multi-scale data structure of ecological environment factors and construction behaviors, the influence of railway tunnel water burst on the environment of the area along the flow path from the view point of the external integral ecological environment of the tunnel water burst for the first time is quantitatively evaluated;
the method comprises the steps of constructing an ecological environment database near a tunnel site area of a mountain railway tunnel in the whole life cycle tunnel construction process by means of a remote sensing satellite image acquisition platform;
the method establishes an evaluation mechanism for the ecological environment of the area along the water flowing path of the railway tunnel for the first time, and fills the blank in the existing ecological environment evaluation research.
The invention constructs and verifies the water flow discharge path of tunnel water burst by combining theoretical deduction and concrete practice, so that the applicability of the unmanned aerial vehicle in the aspect of ecological environment monitoring and evaluation is enhanced while the water flow discharge path is truly simulated;
the method is cut in from the direction of the external whole ecological environment of the water burst of the tunnel, is combined with the actual construction of the railway tunnel, evaluates the influence of the water burst of the railway tunnel on the ecological environment of the area along the flow path at the view angle of the whole life cycle, and accurately and comprehensively reveals the change condition of the influence of the water burst of the tunnel on the ecological environment of the area along the flow path at different periods of the construction of the railway tunnel engineering;
the invention fills the research blank of the railway tunnel water burst ecological environment evaluation, and the proposed tunnel water burst ecological environment risk evaluation model not only inspects the environmental change of the tunnel water burst on each ecological parameter in a specific angle, but also evaluates the ecological environment risk condition of the area along the flow path of the tunnel water burst by constructing the main component containing the most evaluation information, solves the problems that the influence of the existing tunnel water burst on the ecological environment is difficult to evaluate and the quick and accurate real-time dynamic evaluation can not be realized, intuitively displays the influence of the railway tunnel water burst emission on the regional ecological environment, enriches the research result of the railway tunnel engineering ecological environment evaluation, has important significance for the normal construction of the railway tunnel and the protection of the ecological environment around the tunnel site area, the method has a profound effect on achieving the goal of coordinated development of tunnel construction and environmental protection.
According to the invention, before the remote sensing scene images are processed, the remote sensing scene images are preprocessed based on the neural network model, so that the number of the remote sensing scene images is reduced, and the processing efficiency of the system is improved;
for remote sensing scene image classification, many convolutional neural networks improve classification accuracy at the expense of the time and space complexity of the models, which makes the network models difficult to run on mobile devices. Although some existing lightweight convolutional neural networks can provide better classification performance, information interaction among different hierarchical features is not fully considered, so that the improvement of the classification performance is limited;
the target detection network model is an improved U-Net segmentation network based on a traditional U-Net structure, and compared with the traditional U-Net segmentation network, the target detection network model disclosed by the invention has the advantages that two encoders and two decoders are respectively applied in the encoding stage and the decoding stage, and compared with the number of the encoders and the decoders in the traditional U-Net segmentation network, the number of the encoders and the decoders is very small, so that the target detection network model can obtain a better target detection result by using the smaller number of the encoders and the decoders, the calculated amount is reduced, and the target detection efficiency is improved;
the target segmentation network model fully considers the information communication among different levels of features, fully communicates the information among different levels through two different branches, and then the two branches are fused. The method has the advantages that the calculation speed is superior to that of a classification method with the same parameter amount or even less parameter amount, the classification precision is improved, the calculation speed is greatly improved, and the balance of the speed and the precision is realized. In addition, for more efficient feature representation, a combination of max-pooling downsampling and convolutional downsampling is used at the shallow layer of the network for downsampling. Compared with the traditional single down-sampling mode, the down-sampling structure has better performance, and fully considers the information interaction among different hierarchical features, thereby improving the classification performance.
The method comprises the steps of denoising remote sensing image information of a tunnel to be analyzed in three periods before tunnel construction, in a construction period and after the construction; the denoising formula of the invention considers the relation between noise and the information of the remote sensing image of the tunnel to be analyzed, and designs a coefficient of 0.7 by considering various factors involved in the denoising process, so that the image signal obtained after denoising processing is more accurate and smaller, and is not removed more or less, compared with the existing method, the denoising is more accurate, and the problem of incomplete denoising caused by interference of some factors in the existing method is avoided.
Drawings
FIG. 1 is a schematic flow diagram of the present invention;
FIG. 2a is a diagram of an elevation profile of a predetermined area near a case tunnel; FIG. 2b is a diagram showing a slope distribution of the slope of a preset area near the case tunnel; FIG. 2c is a diagram of the distribution of elevation and slope in a preset area near the case tunnel;
FIG. 3a is a case tunnel flow path distribution plot; FIG. 3b is a graph of a case tunnel flow path and 500m profile; FIG. 3c is a layout of a case tunnel flow path and a 1000m buffer area;
FIG. 4a is a 2013 vegetation coverage indicator distribution plot of a case tunnel water inrush flow path along a line buffer zone range; FIG. 4b is a 2017 vegetation coverage indicator distribution plot of a case tunnel water inrush flow path along a line buffer range; FIG. 4c is a 2019 vegetation coverage indicator distribution plot along the line buffer zone of the case tunnel water inrush flow path;
FIG. 4d is a 2013 distribution of humidity indicators along the line buffer range for the case tunnel water inrush flow path; FIG. 4e is a 2017 distribution plot of humidity indicators along the line buffer zone for the case tunnel water inrush flow path; FIG. 4f is a 2019 distribution plot of humidity indicators along the line buffer range for the case tunnel gush flow path;
FIG. 4g is a 2013 dryness fraction indicator distribution plot along the line buffer range for the case tunnel water inrush flow path; FIG. 4h is a 2017 dryness fraction indicator distribution plot of a case tunnel water inrush flow path along a line buffer range; FIG. 4i is a 2019 dryness index distribution plot of a case tunnel water inrush flow path along a line buffer range;
FIG. 4j is a 2013 distribution plot of land surface temperature indicators along the line buffer zone of the case tunnel water flowpath; FIG. 4k is a 2017 distribution plot of land surface temperature indicators along the line buffer zone for the case tunnel water flowpath; FIG. 4l is a 2019 distribution plot of land surface temperature indicators along the line buffer zone of the case tunnel water flowpath;
FIG. 4m is a 2013 salinity indicator profile along the line buffer range for the case tunnel gush water flow path; 4n is 2017 salinity index distribution diagram of the case tunnel water inflow flow path along the line buffer area range; 4o is 2019 salinity index distribution diagram of the case tunnel water inflow flow path along the line buffer area range;
FIG. 5a is a distribution graph of RSEI indicators in 2013 along the line buffer range of the case tunnel water inrush flow path; FIG. 5b is a distribution graph of RSEI indicators in 2017 along the line buffer range of the case tunnel water inrush flow path; fig. 5c is a distribution graph of the RSEI indicators for 2019 along the line buffer range of the case tunnel water inrush flow path.
Detailed Description
The first embodiment is as follows: the ecological environment evaluation system for the railway tunnel water inrush flow path along the line region comprises:
the system comprises a water inrush flow path and buffer area construction module, a multispectral remote sensing image acquisition and preprocessing module, a tunnel water inrush ecological evaluation index construction and ecological index inversion module and a tunnel water inrush flow path region remote sensing ecological improvement model evaluation module;
the water inrush flow path and buffer zone building module is used for determining a flowing path of the water inrush current at the tunnel portal and building a buffer zone along the water inrush flow path by taking the flowing path of the water inrush current at the tunnel portal as a central line;
the multispectral remote sensing image acquisition and preprocessing module is used for acquiring image monitoring data of a tunnel to be analyzed from a remote sensing satellite, and preprocessing the remote sensing image monitoring data to obtain preprocessed multi-period real-time remote sensing image monitoring data of a multi-scale buffer area along a water flowing path of the tunnel to be analyzed, wherein the preprocessed coordinate system is accurate, the cloud cover and the radiation influence are small, and the ground object quality is clear;
the tunnel water burst ecological evaluation index construction and ecological factor index inversion module is used for selecting an ecological evaluation index of a region along a tunnel water flow path according to natural geological conditions of a mountain tunnel site region and by combining construction condition characteristics in a case tunnel construction process; according to the preprocessed multi-period real-time remote sensing image monitoring data of the tunnel multi-scale buffer area to be analyzed, extracting remote sensing ecological factor index parameters required by evaluation and analysis in an inversion mode;
the remote Sensing Ecological improvement model evaluation module of the tunnel water inrush flow path line region is used for constructing an optimal remote Sensing Ecological improvement evaluation model RSEI (remote Sensing Based Ecological index) according to remote Sensing Ecological factor index parameters required by inversion extraction and evaluation, and obtaining the change condition of the Ecological environment of the tunnel water inrush flow path line region before tunnel construction, during construction and after construction Based on the optimal remote Sensing Ecological improvement evaluation model RSEI.
The second embodiment is as follows: the difference between the present embodiment and the first embodiment is that the inrush water flow path and buffer building module is configured to determine a flow path of an inrush water flow at a tunnel entrance, and build an inrush water flow path buffer with the flow path of the inrush water flow at the tunnel entrance as a center line; the specific process is as follows:
the method for determining the flow path of the water flow at the tunnel portal comprises the following steps:
hydrologic analysis is carried out on the research area by means of Google Earth Pro and Arc GIS 10.7 software platforms and by taking DEM data of the area where the tunnel is located as basic data.
Inputting DEM data of the area where the tunnel is located into Arc GIS 10.7 software to obtain hydrological data such as a river network, river length and the like near the area where the tunnel is located, simulating a flowing path of water inrush flow at a tunnel mouth according to the hydrological data such as the river network, the river length and the like of the area where the tunnel is located, and verifying the high-precision image acquired by an unmanned aerial vehicle and a map of the area where the tunnel is located in the past year provided on Google Earth Pro to ensure the accuracy of the image;
construction of a water inflow flow path buffer zone:
according to the flowing path of the water flow of the tunnel entrance, a vector line layer I of the water flow path is established in an Arc GIS 10.7, a multi-scale level buffer area vector range layer II and a multi-scale level buffer area vector range layer III are established by taking the vector line layer I of the water flow path as a central line according to the terrain and feature conditions of the area where the tunnel is located, and buffer area range vector data of the multi-scale level tunnel water flow path are obtained;
the multi-scale buffer is set to a 500 m-scale level buffer and a 1 km-scale level buffer.
Other steps and parameters are the same as those in the first embodiment.
The third concrete implementation mode: the difference between the embodiment and the specific embodiment is that the multispectral remote sensing image acquisition and preprocessing module is used for acquiring image monitoring data of a tunnel to be analyzed from a remote sensing satellite, and preprocessing the remote sensing image monitoring data to obtain preprocessed multi-period real-time remote sensing image monitoring data of a multi-scale buffer area of the tunnel to be analyzed, wherein the preprocessed multi-period real-time remote sensing image monitoring data are accurate in coordinate system, small in cloud amount and radiation influence and clear in ground object quality; the specific process is as follows:
establishing a remote sensing image data set containing a tunnel region: marking a remote sensing image containing a tunnel region in a remote sensing image obtained by a remote sensing satellite to obtain a remote sensing image data set containing the tunnel region;
establishing remote sensing image data sets before tunnel construction, in the construction period and after the construction: marking the obtained remote sensing images which comprise the tunnel region and are collected before tunnel construction, during the construction and after the construction to obtain remote sensing image data sets before tunnel construction, during the construction and after the construction;
establishing a target detection network model;
inputting a remote sensing image data set containing a tunnel region into a target detection network model for training to obtain a trained target detection network model;
establishing a target segmentation network model;
inputting remote sensing image data sets before tunnel construction, during construction and after construction into a target segmentation network model to obtain a trained target segmentation network model;
the method comprises the steps of obtaining a remote sensing image to be interpreted and analyzed from a remote sensing satellite, inputting the remote sensing image to be interpreted and analyzed in a multi-period mode into a trained target detection network model, obtaining the remote sensing image containing a tunnel region, inputting the remote sensing image containing the tunnel region into the trained target segmentation network model, obtaining the remote sensing image of the tunnel to be analyzed before tunnel construction, in a tunnel construction period and after the tunnel construction, preprocessing the remote sensing image of the tunnel to be analyzed in three periods before the tunnel construction, in the construction period and after the tunnel construction period, and obtaining the multi-period real-time remote sensing image monitoring data of the tunnel to be analyzed in a multi-scale buffer area, wherein the preprocessed multi-period real-time remote sensing image monitoring data are accurate in coordinate system, small in cloud amount and radiation influence and clear in ground feature quality.
Other steps and parameters are the same as those in the first or second embodiment.
The fourth concrete implementation mode: the difference between this embodiment and the first to third embodiments is that the target detection network model is established; the specific process is as follows:
the target detection network model adopts a traditional U-Net structure, namely an encoding-decoding (encode-decode) network architecture; two encoders and two decoders are respectively applied in the encoding stage and the decoding stage;
each encoder consists of a feature extraction module and a down-sampling module in sequence;
the feature extraction module is formed by three-layer network cascade, namely a convolution layer formed by convolution kernels with the size of 3 multiplied by 3, the step length of 1 and the padding of 0; secondly, batch normalization layer; thirdly, an active layer formed by a piecewise linear active function ReLU; the downsampling module consists of a single convolution layer, the convolution layer consists of convolution kernels with the size of 3 multiplied by 3 and the step length of 2, and downsampling operation of the feature map is achieved;
each decoder consists of an up-sampling module, a feature splicing module and a feature extraction module in sequence;
the up-sampling module consists of a single deconvolution layer, the deconvolution layer consists of convolution kernels with the size of 2 multiplied by 2, the step length of 2 and the padding of 1, and the up-sampling operation of the characteristic diagram is realized; the feature splicing module splices the feature graph output after up-sampling and the down-sampling feature graph with the same scale in the coding stage on the channel dimension by using a layer-skipping connection mode; the feature extraction module of the decoder is formed by three-layer network cascade, namely a convolution layer formed by convolution kernels with the size of 3 multiplied by 3, the step length of 1 and the padding of 0, a batch normalization layer and an activation layer formed by a piecewise linear activation function ReLU.
Other steps and parameters are the same as those in one of the first to third embodiments.
The fifth concrete implementation mode: the difference between this embodiment and one of the first to the fourth embodiments is that the target segmentation network model is established; the specific process is as follows:
the target segmentation network model comprises an input layer, a first Group1, a second Group2, a third Group3, a fourth Group4, a fifth Group5 and an output classification layer;
the connection relation of the target segmentation network model is as follows:
the output end of the input layer is connected with a first Group1, the output end of a first Group1 is connected with a second Group2, the output end of a second Group2 is connected with a third Group3, the output end of a third Group3 is connected with a fourth Group4, the output end of a fourth Group4 is connected with a fifth Group5, and the output end of a fifth Group5 is connected with the output classification layer to realize classification;
the first Group1 comprises a first max-pooling layer, a first convolution unit, a second convolution unit, a third convolution unit and a first full-link layer;
a second Group2 includes a fourth convolution element, a fifth convolution element, a first depth separable convolutional layer;
a third Group3 includes a second depth separable convolutional layer, a third depth separable convolutional layer, a fourth depth separable convolutional layer, a fifth depth separable convolutional layer, a sixth convolutional unit, a seventh convolutional unit, an eighth convolutional unit, a ninth convolutional unit, a tenth convolutional unit, a second max pooling layer, a first summed add layer, a second summed add layer, a third summed add layer, a fourth summed add layer, a fifth summed add layer, a sixth summed add layer;
a fourth Group4 includes an eleventh convolution element, a twelfth convolution element, and a seventh depth separable convolutional layer;
a fifth Group5 comprises global average pooling GAP, Softmax classification layers.
Other steps and parameters are the same as in one of the first to fourth embodiments.
The sixth specific implementation mode: the difference between this embodiment and one of the first to fifth embodiments is that the output end of the input layer is connected to the input end of the first largest pooling layer in the first Group1 and the input end of the second convolution unit, the output end of the first largest pooling layer is connected to the first convolution unit, the output end of the second largest pooling layer is connected to the third convolution unit, and the output end of the first convolution unit and the output end of the third convolution unit are connected to the first full-connection layer;
the output end of the first full connection layer is connected with the input end of a fourth convolution unit in the second Group2, the output end of the fourth convolution unit is connected with a fifth convolution unit, and the output end of the fifth convolution unit is connected with the first depth separable convolution layer;
the output terminal of the first depth-separable convolutional layer is connected to the input terminal of the second depth-separable convolutional layer, the input terminal of the sixth convolutional unit, and the input terminal of the tenth convolutional unit in the third Group3, respectively;
connecting the output end of the first depth separable convolutional layer, the output end of the second depth separable convolutional layer and the output end of the sixth convolution unit to a first add layer and an add layer, wherein the output ends of the first add layer and the add layer are respectively connected to a third depth separable convolutional layer and a seventh convolution unit;
connecting the output end of the third depth separable convolutional layer, the output end of the sixth convolutional unit, the output end of the seventh convolutional unit and the output end of the first add-add layer to a second add-add layer, and connecting the output ends of the second add-add layer and the add-add layer to a fourth depth separable convolutional layer and an eighth convolutional unit respectively;
connecting the output end of the fourth depth separable convolution layer, the output end of the eighth convolution unit, the output end of the seventh convolution unit, the output end of the sixth convolution unit and the output end of the second add-add layer to the third add-add layer;
connecting the output end of the first depth separable convolutional layer, the output end of the sixth depth separable convolutional layer and the output end of the tenth convolutional unit to a fourth add-add layer and an add layer, wherein the output ends of the fourth add-add layer and the add layer are respectively connected to the fifth depth separable convolutional layer and the ninth convolutional unit;
connecting the output end of the fifth depth separable convolution layer, the output end of the tenth convolution unit, the output end of the ninth convolution unit and the output end of the fourth add-add layer to the fifth add layer and the add layer;
the fifth adding and add layer output end and the third adding and add layer output end are connected with a sixth adding and add layer, and the sixth adding and add layer output end is connected with a second maximum pooling layer;
the output end of the second maximum pooling layer is connected with an eleventh convolution layer in the Group4, the output end of the eleventh convolution layer is connected with a twelfth convolution layer, and the output end of the twelfth convolution layer is connected with a seventh depth separable convolution layer;
the output terminal of the seventh depth separable convolution layer is connected to the global average pooling layer GAP in the fifth Group5, and the global average pooling layer GAP is connected to the Softmax sorting layer for completing the sorting.
Other steps and parameters are the same as those in one of the first to fifth embodiments.
The seventh embodiment: this embodiment is different from one of the first to sixth embodiments in that each of the first to twelfth convolution units includes one convolution layer, one BN layer, and one ReLU layer;
wherein the output end of the convolution layer is connected with the BN layer, and the output end of the BN layer is connected with the ReLU layer;
the convolution kernel size of the convolution layer in the first convolution unit, the second convolution unit, the third convolution unit, the fourth convolution unit, the fifth convolution unit and the twelfth convolution unit is 3 x 3;
the convolution kernel size of the convolution layer in the sixth convolution unit, the seventh convolution unit, the eighth convolution unit, the ninth convolution unit, the tenth convolution unit and the eleventh convolution unit is 1 x 1;
the convolution kernel sizes of the first through seventh depth-separable convolutional layers are 3 × 3.
Other steps and parameters are the same as those in one of the first to sixth embodiments.
The specific implementation mode is eight: the difference between this embodiment and the first to seventh embodiments is that the preprocessing is performed on the remote sensing images of the tunnel to be analyzed in three periods before the tunnel construction, during the tunnel construction and after the tunnel construction, and the specific process is as follows:
the pretreatment method comprises the following steps: denoising, radiometric calibration, geometric correction, atmospheric correction and cutting mosaic;
the remote sensing satellite image is used for acquiring a remote sensing image, an image obtained by the eighth satellite visit of the United states Land satellite plan (Landsat) is adopted, a sensor carried by the image is an OLI (Operational Land Imager), the image acquisition source is (https:// glovis. usgs. gov /), the image precision is 30m, and the tunnel remote sensing image to be analyzed in three periods before tunnel construction, in a tunnel construction period and after tunnel construction is selected.
After the image is obtained, preprocessing the remote sensing image: the method comprises the steps of denoising, radiometric calibration, geometric correction, atmospheric correction, cutting and embedding, registration, image enhancement and the like of the image.
(1) Denoising:
denoising the remote sensing image information of the tunnel to be analyzed in three periods before tunnel construction, in the tunnel construction period and after the tunnel construction;
Figure BDA0003383976140000091
in the formula, f (x, y) represents the remote sensing image information of the tunnel to be analyzed before tunnel construction, in the tunnel construction period and after the tunnel construction, g (x, y) represents the image signal after de-noising processing, and n (x, y) represents noise.
(2) Inputting the information of the remote sensing images of the tunnel to be analyzed in three periods of time before tunnel construction, in the tunnel construction period and after tunnel construction after denoising treatment into ENVI 4.8 to complete radiometric calibration;
radiation calibration: the method aims to eliminate the error of a sensor and determine the accurate radiation value at the inlet of the sensor, and the specific operation in ENVI 4.8 is Basic Tools-Radiometric Correction-Radiometric Calibration;
(3) inputting remote sensing image information of the tunnel to be analyzed in three periods of time before tunnel construction, in the tunnel construction period and after tunnel construction after radiometric calibration into ENVI 4.8 to complete atmospheric correction;
atmospheric correction: aiming at eliminating errors caused by atmospheric scattering, absorption and reflection when the radiation brightness or the surface reflectivity is converted into the actual surface reflectivity, and the specific operation in ENVI 4.8 is Basic colors-preprocessing-calibration functions-FLAASH;
(4) inputting the remote sensing image information of the tunnel to be analyzed in three periods of time before tunnel construction, in the tunnel construction period and after tunnel construction after atmospheric correction into ENVI 4.8 to complete geometric correction;
and (3) geometric correction: the method aims to correct geometric deformation caused by various factors, perform geographic coordinate positioning on the image and obtain real coordinate information. The method selects Orthorectification in geometric correction, and the specific operation in ENVI 4.8 is Map-ortho-rectification;
(5) and cutting and embedding the remote sensing image information of the tunnel to be analyzed in three periods before tunnel construction, in a tunnel construction period and after tunnel construction after geometric correction in Arc GIS 10.7 by taking the multi-scale level buffer area vector range layer II and the layer III as masks to obtain the multi-period real-time remote sensing image monitoring data of the multi-scale buffer area along the water flowing path of the tunnel to be analyzed, which is accurate in coordinate system, small in cloud amount and radiation influence and clear in ground feature quality after pretreatment.
Other steps and parameters are the same as those in one of the first to seventh embodiments.
The specific implementation method nine: the embodiment is different from one to eight of the specific embodiments in that the tunnel water inflow ecological evaluation index construction and ecological factor index inversion module is used for selecting an ecological evaluation index of a region along a tunnel water inflow flow path according to natural geological conditions of a tunnel site region of a mountain tunnel and in combination with a construction condition that the water inflow amount is extremely large in a case tunnel-big pillar mountain tunnel construction process; according to the preprocessed multi-period real-time remote sensing image monitoring data of the tunnel multi-scale buffer area to be analyzed, extracting remote sensing ecological factor index parameters required by evaluation and analysis in an inversion mode; the specific process is as follows:
with the development of the multispectral remote sensing technology, some ecological characteristic parameters can be inverted by using remote sensing image data, and by using the obtained images of the multi-scale level buffer area of the tunnel to be analyzed, according to the natural geological conditions of the area along the tunnel and by combining the construction condition of extremely large water inrush quantity in tunnel construction, the invention selects five remote sensing ecological evaluation indexes, namely vegetation coverage (VFC) reflecting the natural attribute and the greening degree of the area, an improved soil adjustment vegetation index (MSI) based on a salinity index, a dryness index (NDBSI) reflecting the land use change of the area construction and the soil degradation degree, a Land Surface Temperature (LST) and a humidity index (WET), and evaluates the ecological change of the area along the water inrush flow path of the tunnel.
(1) Different vegetation coverage VFC:
the vegetation coverage is the percentage of the vertical projection area of the vegetation (including leaves, stems and branches) on the ground to the total area of the statistical region, is an important index for measuring the coverage condition of the vegetation on the surrounding ground surface of the region along the water flowing path of the tunnel, and can reflect the natural attribute and the greening degree of the region. The vegetation coverage index is calculated on the basis of a normalized vegetation index (NDVI), and the normalized vegetation index is closely related to plant biomass, a leaf area index and vegetation coverage and can reflect the background influence of plant canopy.
ρNDVI＝(ρnir-ρred)/(ρnir+ρred)
In the formula: rhonirThe spectral reflectivity represents the near infrared band of the image; rhoredSpectral reflectance representing the red band of the image; rhoNDVIRepresenting the calculated regional NDVI value, NDVI representing the normalized vegetation index;
after the normalized vegetation index NDVI value of the area is calculated, the band math in ENVI is used for carrying out the following calculation, wherein VFC is equal to VFC
(ρNDVI＞0.7)×1+(ρNDVI＜0.05)×0+(ρNDVI≥0.05&ρNDVI＜0.7)×1(ρNDVI-0.05)/(0.7-0.05)
In the formula: 0.7 represents the NDVI value of the forest land, and 0.05 represents the NDVI value of the bare land;
&representing the median of the NDVI image while satisfying ρNDVINot less than 0.05 and rhoNDVIPixels less than 0.7;
(2) humidity index WET:
the most direct influence of gushing water in the construction and construction process of mountain tunnels on the surrounding ecological environment is vegetation humidity, and the humidity index is calculated by humidity components (WET) in the transform of the tassel caps, so that the humidity index can reflect the humidity of soil and vegetation, is closely related to ecology, and is widely applied to ecological environment monitoring.
WET＝
0.1511ρblue+0.1973ρgreen+0.3283ρred+0.3407ρnir-0.7117ρswir1-0.4559ρswir2
In the formula: wet represents the humidity component of the remote sensing image; rhoblueSpectral reflectance, rho, for the blue band of the imagegreenSpectral reflectance, p, for the green band of the imageredSpectral reflectance, p, for the red band of the imagenirSpectral reflectance, rho, for the near infrared band of the imageswir1Spectral reflectance, rho, for imaging short wave infrared band 1swir2The spectral reflectivity of the image short wave infrared band 2;
(3) dryness index NDBSI:
since human activities are frequent during the construction of the construction area of the tunnel site area, a large amount of bare soil and construction land areas exist, the larger the area of the area represents the more serious land degradation, and the more "dry" area is increased, so that the dryness index (NDSI) is synthesized by both the bare Soil Index (SI) and the construction index (IBI).
SI＝[(ρswir1+ρred)-(ρblue+ρnir)]/[(ρswir1+ρred)+(ρblue+ρnir)]
Figure BDA0003383976140000111
NDBSI＝(SI+IBI)/2
In the formula: SI represents the bare soil index; IBI stands for building index;
(4) improved soil adjustment vegetation index MSI based on salinity index:
the tunnel is inevitably carried with oil stains leaked by tunnel construction, impurities carried by tunnel water burst and other construction wastewater in the water burst water flow after being discharged along the flow path, and the oil stains, the impurities and the other construction wastewater are all important to Soil salinity and Vegetation growth in the peripheral range of the water burst water flow path, so that the salinity is brought into an ecological evaluation Index model of a tunnel water burst area, an improved Soil adjustment Vegetation Index (MSI) based on the salinity Index is established on the basis of an improved Soil adjustment Vegetation Index (MSAIl-Adjusted Vegetation Index), and the optical characteristic change of a Soil background can be better explained, and the sensitivity of NDVI to the Soil background is corrected to a certain extent.
Figure BDA0003383976140000121
Figure BDA0003383976140000122
Wherein SI' represents the salinity index; MSAVI represents the improved soil adjusted vegetation index;
and (3) standardization treatment:
Figure BDA0003383976140000123
Figure BDA0003383976140000124
in the formula: SI'Sign boardRepresents the salinity index after the standardized treatment; MSAIISign boardRepresents the modified soil adjusted vegetation index after standardized treatment; SI'max、SI′minThe maximum value and the minimum value of the salinity index are respectively; MSAIImax、MSAVIminAdjusting the maximum value and the minimum value of the vegetation index for the improved soil respectively;
establishing an improved soil adjustment vegetation index MSI model MSI based on a salinity index:
Figure BDA0003383976140000125
and finally obtaining the improved soil adjustment vegetation index based on the salinity index of the areas near the tunnel water inflow flow path in different years through calculation.
(5) Land surface temperature LST:
considering the complex and dangerous geological condition of the tunnel site area of the mountain tunnel, a large amount of research data show that tunnel water burst is accompanied by high-temperature property, so that the land surface temperature is included in an ecological environment evaluation model along the flow path of the mountain tunnel water burst. At present, the inversion of the surface temperature mainly comprises an atmospheric correction method, a single channel method, a single window algorithm and the like, and the surface temperature (LST) is calculated by adopting the atmospheric correction method. The principle of inverting the landsat-8 surface temperature using atmospheric rectification is: the influence of the atmosphere on the earth surface thermal radiation is firstly estimated, then the part of the atmosphere influence is subtracted from the total amount of the thermal radiation observed by the satellite sensor, so as to obtain the earth surface thermal radiation intensity, and the thermal radiation intensity is converted into the corresponding earth surface temperature.
Carrying out radiometric calibration on the Thermal infrared Thermal band of the image to obtain a radiance image;
calculating normalized vegetation index NDVI
ρNDVI＝(ρnir-ρred)/(ρnir+ρred)
In the formula: rhonirThe spectral reflectivity represents the near infrared band of the image; rhoredSpectral reflectance representing the red band of the image; rhoNDVIRepresenting the calculated regional NDVI value, NDVI representing the normalized vegetation index; rhoNDVIHas a value in the range of-1 to + 1;
calculating vegetation coverage FV:
the vegetation coverage Fv is calculated by adopting a mixed pixel decomposition method, the land category of the whole scene image is roughly divided into a water body, vegetation and a building, and the specific calculation formula is as follows:
FV＝(ρNDVI-ρNDVIS)/(ρNDVIV-ρNDVIS)
the NDVI is a normalized vegetation index, the NDVIV is 0.70, the NDVIS is 0.00, and when the NDVI of a certain pixel is more than 0.70, the FV value is 1; when NDVI is less than 0.00, FV is 0; rhoNDVISRepresenting picture elements of a building, pNDVIVA pixel representing vegetation;
VFC focuses more on the coverage of different vegetation types;
the FV is the vegetation coverage which is calculated after roughly dividing the ground features into water, vegetation and buildings, and is relatively rough, but not as fine as VFC for describing the vegetation.
Inputting by using ENVI main menu- > Basic Tools- > Band Math: (b1 gt 0.7) × 1+ (b1 lt 0.05) × 0+ (b1 ge 0.05 and b1 le 0.7) ((b1-0.05)/(0.7-0.05)), b1 is an NDVI image;
calculating the earth surface emissivity:
dividing the remote sensing image into 3 types of water bodies, towns and natural surfaces according to the characteristics of the area along the tunnel water gushing flow path; the method adopts the following method to calculate the ground surface emissivity of the research area: the emissivity of the water body pixel is assigned to be 0.995, and the emissivity of the natural surface and town pixels is estimated according to the following formula:
εsurface＝0.9625+0.0614×FV-0.0461×FV
εbuilding＝0.9589+0.086×FV-0.0671×FV
in the formula, epsilonsurfaceAnd εbuildingRespectively representing the emissivity of a natural surface pixel and a town pixel;
the formula input column is input by using ENVI main menu- > Basic Tools- > Band Math: (b1 le 0) × 0.995+ (b1 gt 0 and b1 lt 0.7) ((0.9589 +0.086 × b2-0.0671 × b2 × b2) + (b1 ge 0.7) ((0.9625 +0.0614 × b2-0.0461 × b 2b 2)
b 1: NDVI values;
b 2: and (4) vegetation coverage FV value.
Calculating the Black body radiation brightness value Black T at the same temperature:
inquiring the atmospheric profile data (http:// atmcorr. gsfc. nasa. gov /), inputting relevant parameters to obtain atmospheric profile information: transmittance of the atmosphere in the thermal infrared band (t), brightness of the atmosphere radiated upwards (Lu), brightness of the atmosphere radiated downwards (Ld)
Black T＝(ρε-Lu-t×(1-ρε)×Ld)/(t×ρε)
Where rhoεRepresenting the emissivity of the earth's surface, being water, town or natural picture elementsSpecific radiance; lu represents the upward radiation brightness of the atmosphere, and Ld represents the downward radiation brightness of the atmosphere; t represents the transmittance of the atmosphere in the thermal infrared band;
inverting the surface temperature LST:
the following calculation was performed using a band math:
LST＝(K2)/alog(K1/BlackT+1)-273
in the formula, Black T represents a Black body radiation luminance image at the same temperature; k1，K2Is a constant; for Landsat-8 TM sensor, K1 is 774.89W/(m)2·sr·μm)，K2＝1321.08。
Other steps and parameters are the same as those in one to eight of the embodiments.
The detailed implementation mode is ten: the embodiment is different from one of the first embodiment to the ninth embodiment in that the remote sensing ecological improvement model evaluation module for the tunnel water inrush flow path region is used for constructing an optimal remote sensing ecological improvement evaluation model RSEI according to remote sensing ecological factor index parameters required by evaluation extracted by inversion, and obtaining the change conditions of the ecological environment of the tunnel water inrush flow path region before tunnel construction, during construction and after construction based on the optimal remote sensing ecological improvement evaluation model RSEI; the specific process is as follows:
and evaluating a remote sensing ecological improvement model of the tunnel water inrush flow path region according to the vegetation coverage VFC of the region along the tunnel water inrush flow path to be analyzed, the improved soil adjustment vegetation index MSI based on the salinity index, the dryness index NDBSI, the land surface temperature LST and the humidity index WET.
The invention adopts 5 indexes of VFC, Wet, LST, NDBSI and MSI through spatial principal component transformation coupling, objectively determines the weight according to the principal component contribution of different indexes, and realizes the purpose of presenting multi-index information by a single index. However, because the dimensions of different indexes are different, before the principal components are transformed, the dimensions of the indexes need to be unified, the patent normalizes the indexes of different vegetation coverage vfc (Vegetation Fractional coverage), an improved soil adjustment vegetation index MSI based on a salinity index, a dryness index NDBSI, a land surface temperature LST and a humidity index WET, so that the numerical value of each index is between [0 and 1], and the formula is as follows:
RB1＝(b1-bmin)/(bmax-bmin)
in the formula: RB (radio B)1Is the normalized pixel value, b1, bmaxAnd bminRespectively representing the original value, the maximum value and the minimum value of the pixel;
after each ecological factor is synthesized into 1 image, performing principal component analysis based on an ENVI 5.3 platform, and constructing an optimal RSEI according to an analysis result;
the optimal RSEI expression is:
in the RSEI operation process, the RSEI is constructed by using PC1, or by using the contribution ratio of PC1 and PC2 as weight to carry out weighted summation or by using 4 principal components to carry out weighted summation, and the patent sets the principal component contribution ratio to exceed 90% to represent the optimal RSEI for improving the evaluation accuracy. Therefore, the method aims at multiple RSEI construction forms, specifically analyzes specific problems, selects the characteristic components which have the maximum correlation with ecological factors, can reflect the actual situation most and can reasonably explain ecological phenomena in a research area to construct the RSEI.
And combining the ecological factor calculation result and the principal component analysis result to construct a plurality of Remote Sensing Ecological Indexes (RSEIs), carrying out correlation research on the 5 ecological factors and the RSEIs, and exploring the RSEIs with the most comprehensive representativeness in the research area.
Figure BDA0003383976140000151
In the formula, betaiAs contribution ratio of ith principal component, PCiRepresents the ith principal component, i is 1,2,3, 4;
by a PC1Construction of RSEI1From PC1、PC2Weighted summation to construct RSEI2From PC1、PC2、PC3Weighted summation to construct RSEI3From PC1、PC2、PC3、PC4Weighted summation to construct RSEI4；
Based on the operation of the method, the most representative principal component model is screened out through principal component analysis, the optimal RSEI index is calculated, and the most comprehensive representative RSEI evaluation index in the range of the buffer area along the water flow path of the mountain tunnel is obtained.
And classifying the obtained optimal RSEI index by using a natural break point method by means of Arc GIS software, and dividing the optimal RSEI index into four risk grade levels of good, general, poor and extremely poor to obtain the risk conditions of the ecological environment in the area along the water flowing path of the tunnel before tunnel construction, during the construction period and after the construction.
Other steps and parameters are the same as those in one of the first to ninth embodiments.
The following examples were used to demonstrate the beneficial effects of the present invention:
the first embodiment is as follows:
as shown in fig. 1, the remote sensing ecological environment evaluation system for the railway tunnel water inrush flow path along the line region provided by the embodiment includes the following steps:
the method selects a case tunnel in Baoshan city of Yunnan province as an experimental object. The case tunnel passes through a famous southern section of a transverse mountain, the total length is 14484 m, the average altitude is up to 4000 m, the front part of the tunnel is connected with a lanhuajiang bridge and the rear part of the tunnel is connected with a north station of a hilly area, the maximum buried depth of the tunnel is 955m, and the buried depth of a shallow buried section is 55 m. The whole topography north and south is high and low, the geological environment of the region is complex, the fracture structure develops, the geological regions pass through 6 fracture zones, bad geology such as surrounding rock instability, sudden water inrush, rock burst, karst, radioactivity and geothermy can appear in local regions, the construction difficulty is high, and in the actual construction process of the tunnel, the water inrush condition of the tunnel is frequent, and the water inrush quantity is very large.
S01. construction of water inrush flow path and buffer zone
By taking 30m precision DEM data of a preset area of a tunnel address of a case tunnel obtained from a 91-guardian assistant as basic data, simulating an elevation and gradient slope map (shown in figures 2a, 2b and 2 c) of the preset area of the tunnel address of the case tunnel on an Arc GIS software platform, comparing the elevation and gradient slope map with a real-time map on Google Earth, simulating a water flow path of water inrush of the case tunnel according to hydrological data such as a river network, a river length and the like, verifying by using a map of the tunnel address area of the case tunnel in the past year provided by Google Earth Pro and combining field investigation, finding that the water flow path is basically consistent after verification, and having higher accuracy of the simulated water inrush water flow path.
And establishing a vector line map layer of the case tunnel water flow path in Arc GIS 10.7 software, and respectively establishing 500m and 1km buffer area vector ranges by taking the simulated and verified tunnel water flow path as a central line according to the terrain and surface feature conditions of the area where the case tunnel is located to obtain a buffer area vector surface map layer II and a buffer area vector surface map layer III of the multi-scale-level case tunnel water flow path.
S02, acquiring and preprocessing multispectral remote sensing image
The Landsat remote sensing satellite OLI sensor collects remote sensing images of preset areas of the case tunnel in 2013, 2017 and 2019, the remote sensing images to be interpreted and analyzed in three periods before construction of the case tunnel area, in a tunnel construction period and after construction of the tunnel are respectively inspected, the image precision is 30m, and the cloud cover rate is less than 5%. Preprocessing the remote sensing image monitoring data to be interpreted and analyzed in three time periods to obtain the preprocessed multi-period real-time remote sensing image monitoring data of the tunnel multi-scale buffer area to be analyzed, which has the advantages of accurate coordinate system, less influence of cloud amount and radiation and clear ground object quality; the specific process is as follows:
establishing a remote sensing image data set containing a tunnel region: marking a remote sensing image containing a tunnel region in a remote sensing image obtained by a remote sensing satellite to obtain a remote sensing image data set containing the tunnel region;
establishing remote sensing image data sets before tunnel construction, in the construction period and after the construction: marking the obtained remote sensing images which comprise the tunnel region and are collected before tunnel construction, during construction and after construction, and obtaining remote sensing image data sets before tunnel construction, during construction and after construction;
establishing a target detection network model;
inputting a remote sensing image data set containing a tunnel region into a target detection network model for training to obtain a trained target detection network model;
establishing a target segmentation network model;
inputting remote sensing image data sets before tunnel construction, during construction and after construction into a target segmentation network model to obtain a trained target segmentation network model;
inputting remote sensing images of preset areas of case tunnels to be analyzed in 2013, 2017 and 2019 into a trained target detection network model to obtain a remote sensing image containing a tunnel area, inputting the remote sensing image containing the tunnel area into the trained target segmentation network model to obtain remote sensing images of the tunnels to be analyzed before tunnel construction, in a tunnel construction period and after tunnel construction, preprocessing the remote sensing images to be evaluated in three periods of time before construction of the case tunnel area, in the tunnel construction period and after tunnel construction, and obtaining multi-period real-time remote sensing image monitoring data (figures 3a, 3b and 3c) of a multi-scale buffer area of the tunnels to be analyzed, wherein the preprocessed coordinate system is accurate, the cloud amount and radiation influence are small, and the ground quality is clear;
the preprocessing comprises the processing of denoising, radiometric calibration, geometric correction, atmospheric correction, cutting and embedding, registration, image enhancement and the like on the image.
S03, tunnel water burst ecological evaluation index construction and ecological index inversion
By utilizing the image of the multi-scale level buffer area of the case tunnel obtained in the step S02, according to the meteorological geological conditions of the area along the case tunnel and by combining the construction characteristics of frequent water inrush in the tunnel construction process, the invention adopts five remote sensing ecological indexes of vegetation coverage (VFC) reflecting the natural attributes and the greening degree of the area, improved soil adjustment vegetation index (MSI) based on the salinity index, dryness index (NDBSI) reflecting the change of the land for building of the area and the soil degradation degree, and Land Surface Temperature (LST) and humidity index (WET) to evaluate the ecological environment change of the area along the water inrush flow path of the tunnel
Based on an ENVI 5.3 software platform, reflecting the regional natural attribute and the greening degree by using a vegetation coverage VFC, calculating a humidity index by using a humidity component (WET) in the transformation of the tassel cap, synthesizing a dryness index (NDSI) by using a bare Soil Index (SI) and a building index (IBI), and respectively calculating the vegetation coverage, the humidity index and the dryness index by using Band Math to obtain vegetation coverage, humidity index and dryness index distribution maps (4 a, 4b, 4c, 4d, 4e, 4f, 4g, 4h, 4i, 4j, 4k, 4l, 4m, 4n and 4o) of the buffer area along the water inflow flow path of the case tunnel. According to the method, the surface temperature (LST) is calculated by adopting an atmospheric correction method to represent a heat index, and a heat index distribution diagram of the case tunnel water inrush flow path along the buffer area range is obtained (fig. 4a, 4b, 4c, 4d, 4e, 4f, 4g, 4h, 4i, 4j, 4k, 4l, 4m, 4n and 4 o). The variation of the modified soil-adjusted vegetation index (MSI) based on salinity index was obtained by using Band Math inversion to obtain a modified soil-adjusted vegetation index profile based on salinity index along the line buffer range of the tunnel water flow path (fig. 4a, 4b, 4c, 4d, 4e, 4f, 4g, 4h, 4i, 4j, 4k, 4l, 4m, 4n, 4 o).
S04, evaluating remote sensing ecological improvement model of tunnel water burst flow path region
And adjusting the vegetation index distribution diagram according to the vegetation coverage, dryness, heat and humidity of the area of the buffer zone along the case tunnel water flow path obtained in the step S03 and the improved soil based on the salinity index. Because the ecological factor indexes have different dimensions, standardization and normalization are required before combination. After each ecological factor is synthesized into a 1 scene image, principal component analysis is carried out based on an ENVI 5.3 platform, and the RSEI is constructed according to an analysis result.
And combining the ecological factor calculation result and the principal component analysis result to construct a plurality of Remote Sensing Ecological Indexes (RSEIs), carrying out correlation research on the 5 ecological factors and the RSEIs, and exploring the RSEIs with the most comprehensive representativeness in the research area.
Based on the operation of the method, a principal component model which is screened out by principal component analysis and has the most representative is obtained, the optimal RSEI index is calculated, and the RSEI evaluation index which has the most comprehensive representative in the buffer area range along the water inrush flow path of the case tunnel is obtained.
And finally, classifying the obtained optimal RSEI index by using a natural break point method by means of Arc GIS software, and dividing the index into four levels of extreme difference, poor difference, normal and good to obtain RSEI index distribution diagrams (shown in figures 5a, 5b and 5c) in the buffer area range along the tunnel water inflow flow path. Comparing the ecological change conditions of the areas in the buffer area range along the tunnel water flow path on the full life cycle scale before the tunnel construction, during the construction period and after the construction.
It should be noted that the foregoing method embodiments are described as a series of acts or combinations for simplicity in explanation, but it should be understood by those skilled in the art that the present invention is not limited by the order of acts or acts described, as some steps may occur in other orders or concurrently in accordance with the invention.
Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, the patent is not limited to the above-mentioned preferred embodiments, and any other various methods for evaluating the ecological environment of the area along the water flowing path of the railway tunnel can be obtained from the teaching of the patent.
The present invention is capable of other embodiments and its several details are capable of modifications in various obvious respects, all without departing from the spirit and scope of the present invention.

Claims (10)

1. Regional ecological environment evaluation system along line of railway tunnel water flow path that gushes, its characterized in that: the system comprises:
the system comprises a water inrush flow path and buffer area construction module, a multispectral remote sensing image acquisition and preprocessing module, a tunnel water inrush ecological evaluation index construction and ecological index inversion module and a tunnel water inrush flow path region remote sensing ecological improvement model evaluation module;
the water inrush flow path and buffer zone building module is used for determining a flowing path of the water inrush current at the tunnel portal and building a buffer zone along the water inrush flow path by taking the flowing path of the water inrush current at the tunnel portal as a central line;
the multispectral remote sensing image acquisition and preprocessing module is used for acquiring image monitoring data of a tunnel to be analyzed from a remote sensing satellite, and preprocessing the remote sensing image monitoring data to obtain preprocessed multi-period real-time remote sensing image monitoring data of a buffer area along a water inrush flow path of the tunnel to be analyzed;
the tunnel water burst ecological evaluation index construction and ecological factor index inversion module is used for inverting and extracting remote sensing ecological factor index parameters required by evaluation and analysis according to the preprocessed multi-period real-time remote sensing image monitoring data of the tunnel buffer area to be analyzed;
the remote sensing ecological improvement model evaluation module of the tunnel water inrush flow path line region is used for constructing an optimal remote sensing ecological improvement evaluation model RSEI according to remote sensing ecological factor index parameters required by evaluation extracted by inversion, and obtaining the change conditions of the ecological environment of the tunnel water inrush flow path line region before tunnel construction, during construction and after construction based on the optimal remote sensing ecological improvement evaluation model RSEI.
2. The system for evaluating the ecological environment of the area along the railway tunnel water inrush flow path according to claim 1, wherein: the water inrush flow path and buffer zone building module is used for determining a flowing path of the water inrush current at the tunnel entrance and building a water inrush flow path buffer zone by taking the flowing path of the water inrush current at the tunnel entrance as a central line; the specific process is as follows:
the method for determining the flow path of the water flow at the tunnel portal comprises the following steps:
inputting DEM data of the area where the tunnel is located into Arc GIS 10.7 software to obtain a river network and river length data of the area where the tunnel is located, and simulating a flowing path of water flow at the tunnel mouth according to the river network and the river length data of the area where the tunnel is located;
construction of a water inflow flow path buffer zone:
according to the flowing path of the water flow of the tunnel entrance, a vector line layer I of the water flow path is established in Arc GIS 10.7, and a layer II and a layer III are established by taking the vector line layer I of the water flow path as a central line to obtain a buffer area of the water flow path of the tunnel entrance.
3. The system for evaluating the ecological environment of the area along the railway tunnel water inrush flow path according to claim 2, wherein: the multispectral remote sensing image acquisition and preprocessing module is used for acquiring image monitoring data of a tunnel to be analyzed from a remote sensing satellite, and preprocessing the remote sensing image monitoring data to obtain preprocessed multi-period real-time remote sensing image monitoring data of a tunnel buffer area to be analyzed; the specific process is as follows:
establishing a remote sensing image data set containing a tunnel region: marking a remote sensing image containing a tunnel region in a remote sensing image obtained by a remote sensing satellite to obtain a remote sensing image data set containing the tunnel region;
establishing remote sensing image data sets before tunnel construction, in the construction period and after the construction: marking the obtained remote sensing images which comprise the tunnel region and are collected before tunnel construction, during the construction and after the construction to obtain remote sensing image data sets before tunnel construction, during the construction and after the construction;
establishing a target detection network model;
inputting a remote sensing image data set containing a tunnel region into a target detection network model for training to obtain a trained target detection network model;
establishing a target segmentation network model;
inputting remote sensing image data sets before tunnel construction, during construction and after construction into a target segmentation network model to obtain a trained target segmentation network model;
acquiring a remote sensing image to be interpreted and analyzed from a remote sensing satellite, inputting the multi-period remote sensing image to be interpreted and analyzed into a trained target detection network model, acquiring a remote sensing image containing a tunnel region, inputting the remote sensing image containing the tunnel region into the trained target segmentation network model, acquiring the remote sensing image of the tunnel to be analyzed before tunnel construction, in a tunnel construction period and after the tunnel construction, preprocessing the remote sensing image of the tunnel to be analyzed in three periods before the tunnel construction, in the construction period and after the tunnel construction, and acquiring the multi-period real-time remote sensing image monitoring data of the preprocessed tunnel buffer area to be analyzed.
4. The system for evaluating the ecological environment of the area along the railway tunnel water inrush flow path according to claim 3, wherein: establishing a target detection network model; the specific process is as follows:
the target detection network model adopts a traditional U-Net structure, namely an encoding-decoding network architecture; two encoders and two decoders are respectively applied in the encoding stage and the decoding stage;
each encoder consists of a feature extraction module and a down-sampling module in sequence;
the feature extraction module is formed by three-layer network cascade, namely a convolution layer formed by convolution kernels with the size of 3 multiplied by 3, the step length of 1 and the padding of 0; secondly, batch normalization layer; thirdly, an active layer formed by a piecewise linear active function ReLU; the downsampling module consists of a single convolution layer, the convolution layer consists of convolution kernels with the size of 3 multiplied by 3 and the step length of 2, and downsampling operation of the feature map is achieved;
each decoder consists of an up-sampling module, a feature splicing module and a feature extraction module in sequence;
the up-sampling module consists of a single deconvolution layer, the deconvolution layer consists of convolution kernels with the size of 2 multiplied by 2, the step length of 2 and the padding of 1, and the up-sampling operation of the characteristic diagram is realized; the feature splicing module splices the feature graph output after up-sampling and the down-sampling feature graph with the same scale in the coding stage on the channel dimension by using a layer-skipping connection mode; the feature extraction module of the decoder is formed by three-layer network cascade, namely a convolution layer formed by convolution kernels with the size of 3 multiplied by 3, the step length of 1 and the padding of 0, a batch normalization layer and an activation layer formed by a piecewise linear activation function ReLU.
5. The system for evaluating the ecological environment of the area along the railway tunnel water inrush flow path according to claim 4, wherein: establishing a target segmentation network model; the specific process is as follows:
the target segmentation network model comprises an input layer, a first Group1, a second Group2, a third Group3, a fourth Group4, a fifth Group5 and an output classification layer;
the connection relation of the target segmentation network model is as follows:
the output end of the input layer is connected with a first Group1, the output end of a first Group1 is connected with a second Group2, the output end of a second Group2 is connected with a third Group3, the output end of a third Group3 is connected with a fourth Group4, the output end of a fourth Group4 is connected with a fifth Group5, and the output end of a fifth Group5 is connected with the output classification layer to realize classification;
the first Group1 comprises a first max-pooling layer, a first convolution unit, a second convolution unit, a third convolution unit and a first full-link layer;
a second Group2 includes a fourth convolution element, a fifth convolution element, a first depth separable convolutional layer;
a third Group3 includes a second depth separable convolutional layer, a third depth separable convolutional layer, a fourth depth separable convolutional layer, a fifth depth separable convolutional layer, a sixth convolutional unit, a seventh convolutional unit, an eighth convolutional unit, a ninth convolutional unit, a tenth convolutional unit, a second max pooling layer, a first summed add layer, a second summed add layer, a third summed add layer, a fourth summed add layer, a fifth summed add layer, a sixth summed add layer;
a fourth Group4 includes an eleventh convolution element, a twelfth convolution element, and a seventh depth separable convolutional layer;
a fifth Group5 comprises global average pooling GAP, Softmax classification layers.
6. The system for evaluating the ecological environment of the area along the railway tunnel water inrush flow path according to claim 5, wherein: the output end of the input layer is respectively connected with the input end of a first maximum pooling layer in the first Group1 and the input end of a second convolution unit, the output end of the first maximum pooling layer is connected with a first convolution unit, the output end of the second convolution unit is connected with a third convolution unit, and the output end of the first convolution unit and the output end of the third convolution unit are connected with a first full-connection layer;
the output end of the first full connection layer is connected with the input end of a fourth convolution unit in the second Group2, the output end of the fourth convolution unit is connected with a fifth convolution unit, and the output end of the fifth convolution unit is connected with the first depth separable convolution layer;
the output terminal of the first depth-separable convolutional layer is connected to the input terminal of the second depth-separable convolutional layer, the input terminal of the sixth convolutional unit, and the input terminal of the tenth convolutional unit in the third Group3, respectively;
connecting the output end of the first depth separable convolutional layer, the output end of the second depth separable convolutional layer and the output end of the sixth convolution unit to a first add layer and an add layer, wherein the output ends of the first add layer and the add layer are respectively connected to a third depth separable convolutional layer and a seventh convolution unit;
connecting the output end of the third depth separable convolutional layer, the output end of the sixth convolutional unit, the output end of the seventh convolutional unit and the output end of the first add-add layer to a second add-add layer, and connecting the output ends of the second add-add layer and the add-add layer to a fourth depth separable convolutional layer and an eighth convolutional unit respectively;
connecting the output end of the fourth depth separable convolution layer, the output end of the eighth convolution unit, the output end of the seventh convolution unit, the output end of the sixth convolution unit and the output end of the second add-add layer to the third add-add layer;
connecting the output end of the first depth separable convolutional layer, the output end of the sixth depth separable convolutional layer and the output end of the tenth convolutional unit to a fourth add-add layer and an add layer, wherein the output ends of the fourth add-add layer and the add layer are respectively connected to the fifth depth separable convolutional layer and the ninth convolutional unit;
connecting the output end of the fifth depth separable convolution layer, the output end of the tenth convolution unit, the output end of the ninth convolution unit and the output end of the fourth add-add layer to the fifth add layer and the add layer;
the fifth adding and add layer output end and the third adding and add layer output end are connected with a sixth adding and add layer, and the sixth adding and add layer output end is connected with a second maximum pooling layer;
the output end of the second maximum pooling layer is connected with an eleventh convolution layer in the Group4, the output end of the eleventh convolution layer is connected with a twelfth convolution layer, and the output end of the twelfth convolution layer is connected with a seventh depth separable convolution layer;
the output terminal of the seventh depth separable convolution layer is connected to the global average pooling layer GAP in the fifth Group5, and the global average pooling layer GAP is connected to the Softmax sorting layer for completing the sorting.
7. The system of claim 6, wherein the evaluation system of the ecological environment of the area along the railway tunnel water inflow path comprises: each convolution unit in the first convolution unit to the twelfth convolution unit comprises a convolution layer, a BN layer and a ReLU layer;
wherein the output end of the convolution layer is connected with the BN layer, and the output end of the BN layer is connected with the ReLU layer;
the convolution kernel size of the convolution layer in the first convolution unit, the second convolution unit, the third convolution unit, the fourth convolution unit, the fifth convolution unit and the twelfth convolution unit is 3 x 3;
the convolution kernel size of the convolution layer in the sixth convolution unit, the seventh convolution unit, the eighth convolution unit, the ninth convolution unit, the tenth convolution unit and the eleventh convolution unit is 1 x 1;
the convolution kernel sizes of the first through seventh depth-separable convolutional layers are 3 × 3.
8. The system for evaluating the ecological environment of an area along a railway tunnel water inrush flow path according to claim 7, wherein: the method is characterized in that the remote sensing images of the tunnel to be analyzed in three periods before tunnel construction, in the construction period and after the tunnel construction are preprocessed, and the specific process is as follows:
the pretreatment method comprises the following steps: denoising, radiometric calibration, geometric correction, atmospheric correction and cutting mosaic;
(1) denoising:
denoising the remote sensing image information of the tunnel to be analyzed in three periods before tunnel construction, in the tunnel construction period and after the tunnel construction;
Figure FDA0003383976130000051
in the formula, f (x, y) represents the remote sensing image information of the tunnel to be analyzed before construction, in the construction period of the tunnel and after construction of the tunnel, g (x, y) represents the image signal after de-noising processing, and n (x, y) represents noise.
(2) Inputting the information of the remote sensing images of the tunnel to be analyzed in three periods of time before tunnel construction, in the tunnel construction period and after tunnel construction after denoising treatment into ENVI 4.8 to complete radiometric calibration;
(3) inputting remote sensing image information of the tunnel to be analyzed in three periods of time before tunnel construction, in the tunnel construction period and after tunnel construction after radiometric calibration into ENVI 4.8 to complete atmospheric correction;
(4) inputting the remote sensing image information of the tunnel to be analyzed in three periods of time before tunnel construction, in the tunnel construction period and after tunnel construction after atmospheric correction into ENVI 4.8 to complete geometric correction;
(5) and cutting and embedding the remote sensing image information of the tunnel to be analyzed in three periods of time before tunnel construction, in the tunnel construction period and after tunnel construction after geometric correction in Arc GIS 10.7 by taking the layer II and the layer III as masks to obtain the preprocessed multi-period real-time remote sensing image monitoring data of the buffer area along the water flowing path of the tunnel to be analyzed.
9. The system for evaluating the ecological environment of an area along a railway tunnel water inrush flow path according to claim 8, wherein: the tunnel water burst ecological evaluation index construction and ecological factor index inversion module is used for inverting and extracting remote sensing ecological factor index parameters required by evaluation and analysis according to the preprocessed multi-period real-time remote sensing image monitoring data of the tunnel buffer area to be analyzed; the specific process is as follows:
(1) different vegetation coverage VFC:
ρNDVI＝(ρnir-ρred)/(ρnir+ρred)
in the formula: rhonirThe spectral reflectivity represents the near infrared band of the image; rhoredSpectral reflectance representing the red band of the image; rhoNDVIRepresenting the calculated regional NDVI value, NDVI representing the normalized vegetation index;
after the normalized vegetation index NDVI value of the area is calculated, the band math in ENVI is used for the following calculation,
VFC＝
(ρNDVI＞0.7)×1+(ρNDVI＜0.05)×0+(ρNDVI≥0.05&ρNDVI＜0.7)×1(ρNDVI-0.05)/(0.7-0.05)
(2) Humidity index WET:
WET＝
0.1511ρblue+0.1973ρgreen+0.3283ρred+0.3407ρnir-0.7117ρswir1-0.4559ρswir2
in the formula: WET represents the humidity component of the remote sensing image; rhoblueSpectral reflectance, rho, for the blue band of the imagegreenSpectral reflectance, p, for the green band of the imageredSpectral reflectance, p, for the red band of the imagenirSpectral reflectance, rho, for the near infrared band of the imageswir1Spectral reflectance, rho, for imaging short wave infrared band 1swir2The spectral reflectivity of the image short wave infrared band 2;
(3) dryness index NDBSI:
SI＝[(ρswir1+ρred)-(ρblue+ρnir)]/[(ρswir1+ρred)+(ρblue+ρnir)]
Figure FDA0003383976130000061
NDBSI＝(SI+IBI)/2
in the formula: SI represents the bare soil index; IBI stands for building index;
(4) improved soil adjustment vegetation index MSI based on salinity index:
Figure FDA0003383976130000062
Figure FDA0003383976130000063
wherein SI' represents the salinity index; MSAVI represents the improved soil adjusted vegetation index;
and (3) standardization treatment:
Figure FDA0003383976130000064
Figure FDA0003383976130000065
in the formula: SI'Sign boardRepresents the salinity index after the standardized treatment; MSAIISign boardRepresents the modified soil adjusted vegetation index after standardized treatment; SI'max、SI′minThe maximum value and the minimum value of the salinity index are respectively; MSAIImax、MSAVIminAdjusting the maximum value and the minimum value of the vegetation index for the improved soil respectively;
establishing an improved soil adjustment vegetation index MSI model MSI based on a salinity index:
Figure FDA0003383976130000071
(5) land surface temperature LST:
calculating vegetation coverage FV:
the specific calculation formula is as follows:
FV＝(ρNDVI-ρNDVIS)/(ρNDVIV-ρNDVIS)
where ρ isNDVISRepresenting picture elements of a building, pNDVIVA pixel representing vegetation;
calculating the earth surface emissivity:
dividing the remote sensing image into 3 types of water bodies, towns and natural surfaces; the emissivity of the water body pixel is assigned to be 0.995, and the emissivity of the natural surface and town pixels is estimated according to the following formula:
εsurface＝0.9625+0.0614×FV-0.0461×FV
εbuilding＝0.9589+0.086×FV-0.0671×FV
in the formula, epsilonsurfaceAnd εbuildingRespectively representing the emissivity of a natural surface pixel and a town pixel;
calculating the Black body radiation brightness value Black T at the same temperature:
Black T＝(ρε-Lu-t×(1-ρε)×Ld)/(t×ρε)
where rhoεThe emissivity is the emissivity of water, town or natural pixel; lu represents the upward radiation brightness of the atmosphere, and Ld represents the downward radiation brightness of the atmosphere; t represents the transmittance of the atmosphere in the thermal infrared band;
inverting the surface temperature LST:
LST＝(K2)/alog(K1/BlackT+1)-273
in the formula, Black T represents a Black body radiation luminance image at the same temperature; k1，K2Is a constant.
10. The system of claim 9, wherein the evaluation system of the ecological environment of the area along the railway tunnel water inflow path comprises: the tunnel water inrush flow path region remote sensing ecological improvement model evaluation module is used for constructing an optimal remote sensing ecological improvement evaluation model RSEI according to remote sensing ecological factor index parameters required by evaluation extracted by inversion, and obtaining the change conditions of the ecological environment of the tunnel water inrush flow path region before tunnel construction, during construction and after construction based on the optimal remote sensing ecological improvement evaluation model RSEI; the specific process is as follows:
different vegetation coverage VFC, improved soil adjustment vegetation index MSI based on salinity index, dryness index NDBSI, land surface temperature LST and humidity index WET indexes are normalized, so that the numerical value of each index is between [0 and 1], and the formula is as follows:
RB1＝(b1-bmin)/(bmax-bmin)
in the formula: RB (radio B)1Is the normalized pixel value, b1, bmaxAnd bminRespectively representing the original value, the maximum value and the minimum value of the pixel;
after each ecological factor is synthesized into 1 image, performing principal component analysis based on an ENVI 5.3 platform, and constructing an optimal RSEI according to an analysis result;
the optimal RSEI expression is:
Figure FDA0003383976130000081
in the formula, betaiAs contribution ratio of ith principal component, PCiRepresents the ith principal component, i is 1,2,3, 4;
and classifying the obtained optimal RSEI index by using a natural breakpoint method by means of Arc GIS software, wherein the optimal RSEI index is divided into four risk grade levels of good, general, poor and extremely poor, and the risk conditions of the ecological environment of the tunnel water inflow flow path region before the construction of the tunnel, during the construction period and after the construction are obtained.
CN202111445915.XA 2021-11-30 2021-11-30 Ecological environment evaluation system for railway tunnel water inrush flow path along line region Active CN114120147B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111445915.XA CN114120147B (en) 2021-11-30 2021-11-30 Ecological environment evaluation system for railway tunnel water inrush flow path along line region

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111445915.XA CN114120147B (en) 2021-11-30 2021-11-30 Ecological environment evaluation system for railway tunnel water inrush flow path along line region

Publications (2)

Publication Number Publication Date
CN114120147A true CN114120147A (en) 2022-03-01
CN114120147B CN114120147B (en) 2022-12-09

Family

ID=80368592

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111445915.XA Active CN114120147B (en) 2021-11-30 2021-11-30 Ecological environment evaluation system for railway tunnel water inrush flow path along line region

Country Status (1)

Country Link
CN (1) CN114120147B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114781148A (en) * 2022-04-14 2022-07-22 河北地质大学 Surface temperature inversion method and system for thermal infrared remote sensing cloud coverage pixel
CN115063688A (en) * 2022-06-14 2022-09-16 蓝箭航天空间科技股份有限公司 Ecological environment quality assessment method and device
CN115619275A (en) * 2022-10-27 2023-01-17 上海交通大学 Railway engineering ecological restoration decision-making method and device
CN116342556A (en) * 2023-03-30 2023-06-27 中国自然资源航空物探遥感中心 Plateau tunnel potential safety hazard identification method based on thermal infrared remote sensing
CN116757099A (en) * 2023-08-18 2023-09-15 中国科学院南京土壤研究所 Soil salinity inversion and salinization risk assessment method, device and equipment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109581372A (en) * 2018-12-10 2019-04-05 北京航天泰坦科技股份有限公司 A kind of Remote Sensing Monitoring of Ecological Environment method
CN109684929A (en) * 2018-11-23 2019-04-26 中国电建集团成都勘测设计研究院有限公司 Terrestrial plant ECOLOGICAL ENVIRONMENTAL MONITORING method based on multi-sources RS data fusion
CN113553907A (en) * 2021-06-21 2021-10-26 山东建筑大学 Forest ecological environment condition evaluation method based on remote sensing technology
WO2021226976A1 (en) * 2020-05-15 2021-11-18 安徽中科智能感知产业技术研究院有限责任公司 Soil available nutrient inversion method based on deep neural network

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109684929A (en) * 2018-11-23 2019-04-26 中国电建集团成都勘测设计研究院有限公司 Terrestrial plant ECOLOGICAL ENVIRONMENTAL MONITORING method based on multi-sources RS data fusion
CN109581372A (en) * 2018-12-10 2019-04-05 北京航天泰坦科技股份有限公司 A kind of Remote Sensing Monitoring of Ecological Environment method
WO2021226976A1 (en) * 2020-05-15 2021-11-18 安徽中科智能感知产业技术研究院有限责任公司 Soil available nutrient inversion method based on deep neural network
CN113553907A (en) * 2021-06-21 2021-10-26 山东建筑大学 Forest ecological environment condition evaluation method based on remote sensing technology

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
刘瑞: "《川藏铁路特长隧道工程生态影响综合评价》", 《万方数据知识服务平台》 *
聂单南光等: "基于RSEI指数的深圳市生态环境遥感评价", 《江西科学》 *
魏雨涵 等: "《基于 ＲSEI 的漓江流域生态环境质量动态监测》", 《中国水土保持科学》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114781148A (en) * 2022-04-14 2022-07-22 河北地质大学 Surface temperature inversion method and system for thermal infrared remote sensing cloud coverage pixel
CN115063688A (en) * 2022-06-14 2022-09-16 蓝箭航天空间科技股份有限公司 Ecological environment quality assessment method and device
CN115619275A (en) * 2022-10-27 2023-01-17 上海交通大学 Railway engineering ecological restoration decision-making method and device
CN116342556A (en) * 2023-03-30 2023-06-27 中国自然资源航空物探遥感中心 Plateau tunnel potential safety hazard identification method based on thermal infrared remote sensing
CN116757099A (en) * 2023-08-18 2023-09-15 中国科学院南京土壤研究所 Soil salinity inversion and salinization risk assessment method, device and equipment
CN116757099B (en) * 2023-08-18 2024-03-19 中国科学院南京土壤研究所 Soil salinity inversion and salinization risk assessment method, device and equipment

Also Published As

Publication number Publication date
CN114120147B (en) 2022-12-09

Similar Documents

Publication Publication Date Title
CN114120147B (en) Ecological environment evaluation system for railway tunnel water inrush flow path along line region
Guillevic et al. Land Surface Temperature Product Validation Best Practice Protocol Version 1.0-October, 2017
Byrd et al. Evaluation of sensor types and environmental controls on mapping biomass of coastal marsh emergent vegetation
Kamal et al. Assessment of multi-resolution image data for mangrove leaf area index mapping
Yang et al. Urban land-cover change detection through sub-pixel imperviousness mapping using remotely sensed data
Lillesand et al. Remote sensing and image interpretation
Goslee Analyzing remote sensing data in R: the landsat package
Raber et al. Impact of LiDAR nominal post-spacing on DEM accuracy and flood zone delineation
CN101963664B (en) Microwave remote sensing pixel element decomposing method based on land and water living beings classifying information
Rodríguez-López et al. Spectral analysis using LANDSAT images to monitor the chlorophyll-a concentration in Lake Laja in Chile
CN108195767B (en) Estuary wetland foreign species monitoring method
CN111723524B (en) PM2.5 satellite remote sensing inversion method based on daily variation constraint
Liu et al. Comparative analysis of fractional vegetation cover estimation based on multi-sensor data in a semi-arid sandy area
CN114241331B (en) Remote sensing modeling method for ground biomass of reed in wetland by taking UAV as ground and Septinel-2 medium
KR102316598B1 (en) Fabricating system and method of Landcover map, and Program recording media
Tew et al. Urban expansion analysis using Landsat images in Penang, Malaysia
Lou et al. An effective method for canopy chlorophyll content estimation of marsh vegetation based on multiscale remote sensing data
Satti et al. Effects of climate change on vegetation and snow cover area in Gilgit Baltistan using MODIS data
CN110321402B (en) Method for predicting potential distribution of arbor forest in mountainous area
Rodrigues et al. Retrieving total suspended matter in tropical reservoirs within a cascade system with widely differing optical properties
Bao et al. Estimating above-ground biomass of Pinus densata Mast. using best slope temporal segmentation and Landsat time series
Laraby Landsat surface temperature product: global validation and uncertainty estimation
Salman et al. Detection of Spectral Reflective Changes for Temporal Resolution of Land Cover (LC) for Two Different Seasons in central Iraq
Hassan Environmental studies on coastal zone soils of the north Sinai peninsula (Egypt) using remote sensing techniques
Mendoza Remote Sensing Applied to the Study of Environment-Sensitive Chronic Diseases: A Case Study Applied to Quito, Ecuador

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
GR01 Patent grant
GR01 Patent grant