CN111563420A - Sea surface solar flare area oil spilling multispectral detection method based on convolutional neural network - Google Patents
Sea surface solar flare area oil spilling multispectral detection method based on convolutional neural network Download PDFInfo
- Publication number
- CN111563420A CN111563420A CN202010299907.8A CN202010299907A CN111563420A CN 111563420 A CN111563420 A CN 111563420A CN 202010299907 A CN202010299907 A CN 202010299907A CN 111563420 A CN111563420 A CN 111563420A
- Authority
- CN
- China
- Prior art keywords
- neural network
- convolutional neural
- solar flare
- multispectral
- sea surface
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 58
- 238000001514 detection method Methods 0.000 title claims abstract description 34
- 238000000034 method Methods 0.000 claims abstract description 39
- 239000003305 oil spill Substances 0.000 claims abstract description 25
- 239000011159 matrix material Substances 0.000 claims abstract description 24
- 238000001228 spectrum Methods 0.000 claims abstract description 22
- 238000000605 extraction Methods 0.000 claims abstract description 9
- 230000004913 activation Effects 0.000 claims description 31
- 239000013598 vector Substances 0.000 claims description 18
- 230000003595 spectral effect Effects 0.000 claims description 15
- 210000002569 neuron Anatomy 0.000 claims description 10
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 7
- 230000009467 reduction Effects 0.000 claims description 5
- 238000007634 remodeling Methods 0.000 claims description 4
- 238000003672 processing method Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 description 34
- 239000010410 layer Substances 0.000 description 33
- 238000010586 diagram Methods 0.000 description 12
- 238000012545 processing Methods 0.000 description 10
- 238000004590 computer program Methods 0.000 description 8
- 238000002474 experimental method Methods 0.000 description 7
- 238000012706 support-vector machine Methods 0.000 description 6
- 238000007637 random forest analysis Methods 0.000 description 5
- 238000007476 Maximum Likelihood Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000000691 measurement method Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 241000251468 Actinopterygii Species 0.000 description 1
- 125000003118 aryl group Chemical group 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012567 pattern recognition method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000682 scanning probe acoustic microscopy Methods 0.000 description 1
- 239000013545 self-assembled monolayer Substances 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Multimedia (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a multispectral detection method for oil spilling in sea surface solar flare areas based on a convolutional neural network, which comprises the following steps: denoising multispectral data of an oil film image of a sea surface solar flare area to obtain denoised data; converting the de-noising data into a two-dimensional spectrum matrix; inputting the two-dimensional spectrum matrix into a convolutional neural network as input data; and performing feature extraction and classification by using the convolutional neural network, and outputting a classification result. According to the method, the sea surface solar flare area oil spill extraction model based on the Convolutional Neural Network (CNN) is constructed, the oil film of the sea surface solar flare area is extracted, the convolutional neural network model has higher precision and consistency for classification of the sea surface solar flare area oil spill, and the deep information of the oil spill image can be automatically mined by the convolutional neural network due to the characteristics of local connection, weight sharing and the like, so that more and more essential characteristics are learned, and the optimal classification accuracy is obtained.
Description
Technical Field
The invention relates to the field of ocean detection, in particular to a sea surface solar flare area oil spilling multispectral detection method based on a convolutional neural network.
Background
According to data statistics of the international oil tanker owner pollution alliance (ITOPF), 358 times of large and medium-sized oil spill accidents caused about 113 million tons of oil leakage in the 90 s of the 20 th century. In the beginning of the 20 th century, there were 181 large and medium oil spill accidents, resulting in 19.6 million tons of oil leakage. In the nine years from 2010 to 2018, 59 large and medium oil spillage accidents occur, and 16.3 million tons of oil are leaked. Sea surface oil spillage can pollute the ocean and seriously destroy the ocean ecology. In 4 months in 2010, the oil platform in the deepwater horizon in the gulf of mexico leaks, which pollutes the coastline of more than 160 kilometers in louisiana, destroys beaches, kills a large amount of fish, and destroys the entire ecological system in the gulf of mexico. In 2011, 6 months, oil leakage in oil fields 19-3 Penglai province in Shandong province pollutes the surrounding oil fields and the sea area in northwest of the oil fields.
The offshore oil spill detection mainly takes satellite remote sensing as a main part and mainly comprises two means of microwave remote sensing and multispectral remote sensing. When a large-area oil spill accident occurs on the sea surface, due to the existence of the solar flare area, the oil film of the spilled oil on the sea surface can generate light and shade changes on the remote sensing image, the classification of the remote sensing image can be seriously interfered, and the problem of accurately detecting the spilled oil in the solar flare area on the sea surface is the current difficulty of detecting the oil spill on the sea surface.
In order to solve the problem that the oil spill remote sensing detection is influenced by noise, some researches firstly carry out image filtering and then feature extraction, and then classify the oil spill by using the features. Many researchers use traditional pattern recognition methods such as Support Vector Machines (SVMs), Spectral Angle Matching (SAMs), Artificial Neural Network (ANN) algorithms, etc. For example, the oil film extraction is carried out by using a spectral angle matching method of auxiliary texture characteristic quantity such as the sun-yuan aromatic, and the like, and the precision is up to more than 90%. Calabres et al propose a neural network method for semi-automatic detection of oil spill in ERS-SAR images. However, the accuracy of the existing identification method is difficult to further improve, and the method falls into a bottleneck.
In view of the above, it is difficult for the detecting party in the prior art to further improve the accuracy, and therefore, a method capable of improving the detection accuracy is urgently needed.
Disclosure of Invention
In view of this, the invention provides a method for multispectral detection of oil spilling in a sea surface solar flare area based on a convolutional neural network, so as to solve the problem that the detection accuracy is difficult to improve in the measurement method in the prior art.
In order to achieve the purpose, the technical scheme of the multispectral detection method for oil spilling in the sea surface solar flare area based on the convolutional neural network provided by the invention is as follows:
a sea surface solar flare area oil spilling multispectral detection method based on a convolutional neural network, the detection method comprising:
denoising multispectral data of an oil film image of a sea surface solar flare area to obtain denoised data;
converting the de-noising data into a two-dimensional spectrum matrix;
inputting the two-dimensional spectrum matrix into a convolutional neural network as input data;
and performing feature extraction and classification by using the convolutional neural network, and outputting a classification result.
Preferably, the denoising processing method is to perform filtering processing on the multispectral data by using a median filter.
Preferably, the method for converting the de-noised data into a two-dimensional spectrum matrix comprises the following steps:
performing matrix remodeling and dimension reduction operation on the de-noising data to obtain a spectral vector;
performing band operation on the spectrum vector to obtain an expanded spectrum vector;
and performing dimension raising operation on the expanded spectrum vector to obtain the two-dimensional spectrum matrix.
Preferably, the convolutional neural network comprises an input layer, a first convolutional layer, a second convolutional layer, a fully-connected layer and an output layer.
Preferably, the convolution kernel sizes of the first convolution layer and the second convolution layer are both 3 × 3 and all zero padding is used.
Preferably, the neurons of the second convolutional layer have a first activation function running thereon.
Preferably, the first activation function is a Sigmoid activation function.
Preferably, a second activation function is run on the neurons of the fully connected layer.
Preferably, the second activation function is a Sigmoid activation function.
Preferably, the classification result output by the output layer includes three data labels, which respectively represent the water body, the dark oil film and the bright oil film.
The multispectral detection method for oil spilling in sea surface solar flare areas based on the convolutional neural network has the beneficial effects that:
according to the method, the sea surface solar flare area oil spill extraction model based on the Convolutional Neural Network (CNN) is constructed, the oil film of the sea surface solar flare area is extracted, the convolutional neural network model has higher precision and consistency for classification of the sea surface solar flare area oil spill, and the deep information of the oil spill image can be automatically mined by the convolutional neural network due to the characteristics of local connection, weight sharing and the like, so that more and more essential characteristics are learned, and the optimal classification accuracy is obtained.
Drawings
The drawings are included to provide a better understanding of the invention and are not to be construed as unduly limiting the invention. Wherein:
fig. 1 is a flowchart of a method for multispectral detection of oil spill of a sea surface solar flare area based on a convolutional neural network according to an embodiment of the present invention;
fig. 2 is a structural diagram of a convolutional neural network in a sea surface solar flare area oil spilling multispectral detection method based on the convolutional neural network according to an embodiment of the present invention;
FIG. 3 is a pseudo-color image of two experimental regions in a detection method according to an embodiment of the present invention;
FIG. 4 is a diagram of the classification result of the experimental area 1 by using the method provided by the present invention and the classification of the support vector machine, the random forest classification and the maximum likelihood classification;
FIG. 5 is a diagram of the classification result of the experimental area 2 by using the method provided by the present invention and the classification of the support vector machine, the random forest classification and the maximum likelihood classification.
Detailed Description
The present invention will be further described with reference to the following examples.
Aiming at the problem that the detection precision is difficult to improve in the measurement method in the prior art, the applicant finds that the convolutional neural network has a strong framework for directly learning image feature expression from mass image data, and the classification precision is expected to be improved if the convolutional neural network is applied to classification of oil spilling images in solar flare areas. Based on this, the present embodiment provides a method for multispectral detection of oil spill from a sea surface solar flare area based on a convolutional neural network, as shown in fig. 1, the method for detection includes:
s100, denoising multispectral data of an oil film image in a sea surface solar speckle region to obtain denoising data;
s200, converting the de-noising data into a two-dimensional spectrum matrix;
s300, inputting the two-dimensional spectrum matrix into a convolutional neural network as input data;
and S400, performing feature extraction and classification by using the convolutional neural network, and outputting a classification result.
In S100, multispectral characteristic data of the oil film image of the sea surface solar flare area are adopted for processing to detect.
Since the solar flare phenomenon in the image significantly affects the classification accuracy, in the present application, the denoising data is obtained by performing denoising processing on the multispectral data in step S100. For example, in a specific embodiment, a Median Filter (media Filter) is used to Filter the multispectral data, so as to effectively suppress the influence of solar flare on the detection accuracy.
The data form of the multispectral data cannot be directly used as input data of a convolutional neural network for processing, and aiming at the problem, denoising data are firstly converted into a two-dimensional spectral matrix, and then the two-dimensional spectral matrix is used as the input data and input into the convolutional neural network for operation.
In a specific embodiment, the method for converting the de-noised data into a two-dimensional spectrum matrix comprises the following steps:
performing matrix remodeling and dimension reduction operation on the de-noising data to obtain a spectral vector;
performing band operation on the spectrum vector to obtain an expanded spectrum vector;
and performing dimension raising operation on the expanded spectrum vector to obtain the two-dimensional spectrum matrix.
The structure of the convolutional neural network can be set according to specific requirements by comparing the classification accuracy of different structural models, and in a preferred embodiment, as shown in fig. 2, the convolutional neural network includes an input layer, a first convolutional layer, a second convolutional layer, a fully-connected layer, and an output layer. Wherein the convolution kernel sizes of the first convolution layer and the second convolution layer are both 3 x 3 and all zero padding is used.
After convolution, an activation function is preferably introduced, namely, a first activation function runs on neurons of the second convolution layer, and further, a nonlinear activation function (activation function) is preferably introduced for activation. The first activation function is, for example, a Relu activation function, a Sigmoid activation function, or the like, and preferably a Sigmoid activation function is used.
Similarly, an activation function is also operated on the neurons of the full connection layer, and is marked as a second activation function, and the second activation function also preferably adopts a Sigmoid activation function.
In step S400, the output classification result may be set as required, and preferably, the classification result includes three data tags respectively representing the water body, the dark oil film, and the bright oil film.
A specific embodiment of the detection method provided in the present application is given below.
1. Data processing
The data used in this example are Lansat7 ETM + remote sensing image data of 5 months and 1 day 2010. Two experimental areas (respectively marked as an experimental area 1 and an experimental area 2) are selected in the experiment, the sizes of the image experimental areas are respectively 200 × 302 and 200 × 300, and the total number of the bands is 7. FIG. 3 is a pseudo color composite of the multi-spectral images of experiment 1 and experiment 2, with the 3 channels red, green and blue corresponding to 662.0nm, 560.0nm and 483.0nm, respectively.
The solar flare phenomenon of the image in the experimental area obviously affects the classification precision seriously, so the Median Filter (media Filter) with the size of 3 × 3 of the template is selected for filtering the whole image, thereby inhibiting the influence of the solar flare on the experimental result.
Further, in this embodiment, before the experiment, the three-dimensional experimental data is subjected to the dimension reduction processing, and is converted into the two-dimensional spectral matrix which can be recognized and read by the convolutional neural network model.
The data of the two experimental areas are remote sensing images of 302 × 200 × 6 and 300 × 200 × 6 respectively, spectral information of each pixel in the images is converted into spectral vectors of 1 × 6 dimensions through matrix remodeling and dimension reduction operations, in order to adapt to input of a convolutional neural network, the spectral vectors are respectively subjected to band operation and expanded into spectral vectors of 60400 × 144 and 60000 × 144, and then the multispectral images are converted into spectral matrices of 60400 × 12 × 12 by using dimension increasing operations and serve as input data of input layer neurons of a convolutional neural network model.
2. Construction of sea surface solar flare area oil spill detection model based on convolutional neural network
The structure of the convolutional neural network is as described above.
Wherein the input layer is used for receiving input data, and the input data is a spectral matrix of 12 × 12 × 60400, so the size of the input layer of the convolutional neural network is set to 12 × 12.
The convolution layer is used for extracting different input features, and convolution is an effective image feature extraction method. Typically, a square convolution kernel is used to traverse each pixel in the image. And multiplying the weight of the corresponding point in the convolution kernel by the value of each pixel in the overlapping area of the image and the convolution kernel, summing the values and adding the offset to obtain the pixel value in the output image, wherein the specific formula is as follows:
where p and q represent the height and width, w, of the input convolution kernel, respectivelyiRepresents a weight, viRepresenting the value of each pixel, and b is the offset.
After convolution, a nonlinear activation function (activation function) is introduced for activation, and mainstream nonlinear activation functions include a Relu activation function, a Sigmoid activation function and the like. The Sigmoid function has the advantages that (1) the smooth function (2) convenient for derivation can compress data, the data amplitude is guaranteed not to have the problem (3) and the Sigmoid function is suitable for forward propagation, and the like, so that the Sigmoid function is adopted in the embodiment, and the formula is as follows:
where x is the input characteristic, e is the natural base number, and h (x) is the output of the function. The convolution kernels of the first convolution layer Conv1 and the second convolution layer Conv2 are both 3 × 3 in size and all-zero padding is used, the number of signatures output by the first convolution layer Conv1 is 5, and the number of signatures output by the second convolution layer Conv2 is 7.
Regarding the full-connected layer, FeatureMaps2 are connected into one (8 × 8) × 7 ═ 448 vector, in this embodiment, a 50-sample batch training method is adopted, so FeatureMaps2 are spliced into one 448 × 50 feature vector Fv, the feature vector Fv is used as the input of the single-layer perceptron, the output layer is obtained in a full-connected manner, and the final output class number is set to be 3, that is, three classes of ground objects, namely, a water body, a dark oil film and a bright oil film.
Each neuron in the fully connected layer (FC) is fully connected to all neurons in the layer before it, which can integrate discriminative information extracted in the convolutional layer for final classification. Each neuron in the fully-connected layer is also activated by a stimulus function, which is a sigmoid function in this embodiment.
Further, in the present embodiment, a logistic regression (softmax regression) is used for classification, and the formula is as follows:
in the formula yiFor input features, e is a natural base number and n is an output class number.
In order to verify the effectiveness of the method of the present invention, the method is compared with three methods, namely Support Vector Machine Classification (SVM), Maximum likelihood Classification (ML) and Random forest Classification (RF) which are used for oil overflow Classification in recent literature. The classification experiment is carried out on the bright oil film, the dark oil film and the water body in the experiment area under the same experiment condition,
tables 1 and 2 show confusion matrices for the classification results of the convolutional neural network models.
TABLE 1 Experimental area 1 convolutional neural network model confusion matrix
TABLE 2 Experimental area 2 convolutional neural network model confusion matrix
The overall classification accuracy OA, quality factor QF, Kappa coefficient of the respective classification methods were obtained by calculation, as shown in tables 3 and 4.
TABLE 3 comparison of results of different classification methods in Experimental zone 1
TABLE 4 comparison of results of different classification methods in Experimental section 2
As can be seen from the data in tables 3 and 4, the overall classification accuracies OA of the convolutional neural network models in the two experimental areas are respectively 99.97% and 95.89%, the Kappa coefficients are respectively 1 and 0.92, the quality factors are respectively 1 and 0.98, and the overall classification accuracies OA are higher than those of the other three classification methods, so that the convolutional neural network models constructed by the method have higher accuracy and consistency for classification of the sea surface solar flare area oil spill.
Fig. 4 and 5 respectively show the classification result diagrams of experimental areas 1 and 2 by using the method provided by the invention and the classification of the support vector machine, the random forest classification and the maximum likelihood classification, and as can be seen from fig. 4 and 5, the model of the invention has good classification effect on the oil film under the solar flare area, the precision can reach 95% -100%, and the influence of solar flare on the oil film classification can be effectively avoided. The model has high detection precision on the oil film edge, avoids the misclassification of the oil film and the water body to a certain extent, and the other three classification methods are relatively rough in performance and generate the misclassification of part of the water body and the oil film. The method shows that the characteristics of local connection, weight sharing and the like of the convolutional neural network model enable the convolutional neural network model to automatically mine deep information of oil spilling, learn more and more essential characteristics and further obtain better classification accuracy.
As will be appreciated by one skilled in the art, embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present disclosure, and it should be understood by those skilled in the art that various modifications or variations may be made without inventive faculty based on the technical solutions of the present invention.
Claims (10)
1. A sea surface solar flare area oil spilling multispectral detection method based on a convolutional neural network is characterized by comprising the following steps:
denoising multispectral data of an oil film image of a sea surface solar flare area to obtain denoised data;
converting the de-noising data into a two-dimensional spectrum matrix;
inputting the two-dimensional spectrum matrix into a convolutional neural network as input data;
and performing feature extraction and classification by using the convolutional neural network, and outputting a classification result.
2. The method according to claim 1, wherein the denoising processing method is to filter the multispectral data by using a median filter.
3. The method for multispectral detection of sea surface solar flare area oil spill based on a convolutional neural network as claimed in claim 1, wherein the method for converting the de-noised data into a two-dimensional spectral matrix comprises:
performing matrix remodeling and dimension reduction operation on the de-noising data to obtain a spectral vector;
performing band operation on the spectrum vector to obtain an expanded spectrum vector;
and performing dimension raising operation on the expanded spectrum vector to obtain the two-dimensional spectrum matrix.
4. The method for multispectral detection of sea surface solar flare spill oil based on a convolutional neural network as claimed in any one of claims 1 to 3, wherein the convolutional neural network comprises an input layer, a first convolutional layer, a second convolutional layer, a fully connected layer and an output layer.
5. The convolutional neural network based sea surface solar flare region oil spill multispectral detection method of claim 4, wherein the convolution kernel sizes of the first convolutional layer and the second convolutional layer are both 3 x 3 and all zero padding is used.
6. The convolutional neural network based multispectral sea solar flare region oil spill detection method of claim 4, wherein a first activation function is run on neurons of the second convolutional layer.
7. The convolutional neural network based multispectral sea solar flare region oil spill detection method of claim 6, wherein the first activation function is a Sigmoid activation function.
8. The convolutional neural network based multispectral sea solar flare region oil spill detection method of claim 4, wherein a second activation function is run on neurons of the fully connected layer.
9. The convolutional neural network based multispectral sea solar flare area oil spill detection method of claim 8, wherein the second activation function is a Sigmoid activation function.
10. The method for multispectral detection of sea surface solar flare area oil spill based on a convolutional neural network as claimed in claim 4, wherein the classification result output by the output layer comprises three data labels respectively representing a water body, a dark oil film and a bright oil film.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010299907.8A CN111563420A (en) | 2020-04-16 | 2020-04-16 | Sea surface solar flare area oil spilling multispectral detection method based on convolutional neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010299907.8A CN111563420A (en) | 2020-04-16 | 2020-04-16 | Sea surface solar flare area oil spilling multispectral detection method based on convolutional neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111563420A true CN111563420A (en) | 2020-08-21 |
Family
ID=72074356
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010299907.8A Pending CN111563420A (en) | 2020-04-16 | 2020-04-16 | Sea surface solar flare area oil spilling multispectral detection method based on convolutional neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111563420A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112558187A (en) * | 2020-12-17 | 2021-03-26 | 昆明理工大学 | Method for forecasting solar flare outbreak based on 3D convolutional neural network |
CN113505712A (en) * | 2021-07-16 | 2021-10-15 | 自然资源部第一海洋研究所 | Novel loss function based sea surface oil spill detection method of convolutional neural network |
CN113610762A (en) * | 2021-07-07 | 2021-11-05 | 中国科学院国家空间科学中心 | Early warning method and early warning system for solar flare |
CN114511790A (en) * | 2022-02-11 | 2022-05-17 | 广东海启星海洋科技有限公司 | Method and device for realizing marine oil spill event monitoring based on optical satellite data |
CN114539586A (en) * | 2022-04-27 | 2022-05-27 | 河南银金达新材料股份有限公司 | Surface treatment production process of polymer film |
CN114863293A (en) * | 2022-05-07 | 2022-08-05 | 中国石油大学(华东) | Hyperspectral oil spill detection method based on double-branch GAN network |
CN116206221A (en) * | 2023-01-16 | 2023-06-02 | 北京师范大学 | Water flare detection method and system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107292343A (en) * | 2017-06-23 | 2017-10-24 | 中南大学 | A kind of Classification of hyperspectral remote sensing image method based on six layers of convolutional neural networks and spectral space information consolidation |
CN108334937A (en) * | 2018-02-06 | 2018-07-27 | 大连海事大学 | A kind of oil film relative thickness extracting method and system waterborne based on convolutional neural networks |
CN109919123A (en) * | 2019-03-19 | 2019-06-21 | 自然资源部第一海洋研究所 | Offshore spilled oil detection method based on Analysis On Multi-scale Features depth convolutional neural networks |
CN110348456A (en) * | 2019-04-30 | 2019-10-18 | 浙江大学 | A kind of floating HNS object detection method of combination multispectral image and deep learning method |
-
2020
- 2020-04-16 CN CN202010299907.8A patent/CN111563420A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107292343A (en) * | 2017-06-23 | 2017-10-24 | 中南大学 | A kind of Classification of hyperspectral remote sensing image method based on six layers of convolutional neural networks and spectral space information consolidation |
CN108334937A (en) * | 2018-02-06 | 2018-07-27 | 大连海事大学 | A kind of oil film relative thickness extracting method and system waterborne based on convolutional neural networks |
CN109919123A (en) * | 2019-03-19 | 2019-06-21 | 自然资源部第一海洋研究所 | Offshore spilled oil detection method based on Analysis On Multi-scale Features depth convolutional neural networks |
CN110348456A (en) * | 2019-04-30 | 2019-10-18 | 浙江大学 | A kind of floating HNS object detection method of combination multispectral image and deep learning method |
Non-Patent Citations (3)
Title |
---|
YU HUI ET AL.: "Oil spill detection using hyperspectral infrared camera", 《PROC. SPIE 10030,INFRARED, MILLIMETER-WAVE, AND TERAHERTZ TECHNOLOGIES IV》 * |
安超 等: "太阳耀斑图像中油膜呈现明暗变化原因分析", 《海洋湖沼通报》 * |
王璨: "基于机器学习的农业图像识别与光谱检测方法研究", 《中国博士学位论文全文数据库 农业科技辑》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112558187A (en) * | 2020-12-17 | 2021-03-26 | 昆明理工大学 | Method for forecasting solar flare outbreak based on 3D convolutional neural network |
CN112558187B (en) * | 2020-12-17 | 2022-08-12 | 昆明理工大学 | Method for forecasting solar flare outbreak based on 3D convolutional neural network |
CN113610762A (en) * | 2021-07-07 | 2021-11-05 | 中国科学院国家空间科学中心 | Early warning method and early warning system for solar flare |
CN113610762B (en) * | 2021-07-07 | 2024-02-23 | 中国科学院国家空间科学中心 | Early warning method and early warning system for solar flare |
CN113505712A (en) * | 2021-07-16 | 2021-10-15 | 自然资源部第一海洋研究所 | Novel loss function based sea surface oil spill detection method of convolutional neural network |
CN113505712B (en) * | 2021-07-16 | 2024-06-11 | 自然资源部第一海洋研究所 | Sea surface oil spill detection method of convolutional neural network based on quasi-balance loss function |
CN114511790A (en) * | 2022-02-11 | 2022-05-17 | 广东海启星海洋科技有限公司 | Method and device for realizing marine oil spill event monitoring based on optical satellite data |
CN114539586A (en) * | 2022-04-27 | 2022-05-27 | 河南银金达新材料股份有限公司 | Surface treatment production process of polymer film |
CN114539586B (en) * | 2022-04-27 | 2022-07-19 | 河南银金达新材料股份有限公司 | Surface treatment production and detection process of polymer film |
CN114863293A (en) * | 2022-05-07 | 2022-08-05 | 中国石油大学(华东) | Hyperspectral oil spill detection method based on double-branch GAN network |
CN116206221A (en) * | 2023-01-16 | 2023-06-02 | 北京师范大学 | Water flare detection method and system |
CN116206221B (en) * | 2023-01-16 | 2023-09-05 | 北京师范大学 | Water flare detection method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111563420A (en) | Sea surface solar flare area oil spilling multispectral detection method based on convolutional neural network | |
Han et al. | Underwater image processing and object detection based on deep CNN method | |
Rathi et al. | Underwater fish species classification using convolutional neural network and deep learning | |
Ali et al. | Structural crack detection using deep convolutional neural networks | |
Shamsolmoali et al. | A novel deep structure U-Net for sea-land segmentation in remote sensing images | |
Li et al. | Multiscale features supported DeepLabV3+ optimization scheme for accurate water semantic segmentation | |
Cui et al. | SANet: A sea–land segmentation network via adaptive multiscale feature learning | |
Xu et al. | Fast method of detecting tomatoes in a complex scene for picking robots | |
CN108427920A (en) | A kind of land and sea border defense object detection method based on deep learning | |
CN110120020A (en) | A kind of SAR image denoising method based on multiple dimensioned empty residual error attention network | |
CN109086824A (en) | A kind of sediment sonar image classification method based on convolutional neural networks | |
CN104299006A (en) | Vehicle license plate recognition method based on deep neural network | |
CN106650812A (en) | City water body extraction method for satellite remote sensing image | |
CN109785344A (en) | The remote sensing image segmentation method of binary channel residual error network based on feature recalibration | |
CN111862143B (en) | Automatic monitoring method for river dike collapse | |
CN109410228A (en) | Internal wave of ocean detection algorithm based on Method Based on Multi-Scale Mathematical Morphology Fusion Features | |
Chen et al. | A rotation-invariant convolutional neural network for image enhancement forensics | |
CN111046772A (en) | Multi-temporal satellite remote sensing island shore line and development and utilization information extraction method | |
CN113139550B (en) | Remote sensing image coastline extraction method based on deep semantic segmentation network | |
CN111488940A (en) | Intelligent navigation mark image classification method based on deep convolutional neural network | |
Li et al. | Densely connected network for impulse noise removal | |
CN111553227A (en) | Lightweight face detection method based on task guidance | |
CN108765439A (en) | A kind of sea horizon detection method based on unmanned water surface ship | |
CN117315446B (en) | Reservoir spillway abnormity intelligent identification method oriented to complex environment | |
Yu et al. | Deep multi-feature learning for water body extraction from Landsat imagery |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200821 |
|
RJ01 | Rejection of invention patent application after publication |