CN115063596A - Water quality turbidity detection method based on deep regression network - Google Patents
Water quality turbidity detection method based on deep regression network Download PDFInfo
- Publication number
- CN115063596A CN115063596A CN202210683908.1A CN202210683908A CN115063596A CN 115063596 A CN115063596 A CN 115063596A CN 202210683908 A CN202210683908 A CN 202210683908A CN 115063596 A CN115063596 A CN 115063596A
- Authority
- CN
- China
- Prior art keywords
- turbidity
- network
- image
- regression network
- deep regression
- 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
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 43
- 238000001514 detection method Methods 0.000 title claims abstract description 19
- 238000000034 method Methods 0.000 claims abstract description 13
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims abstract description 8
- 238000013507 mapping Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 6
- 238000011478 gradient descent method Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000004069 differentiation Effects 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 description 5
- 238000003384 imaging method Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012634 optical imaging Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/05—Underwater scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/70—Labelling scene content, e.g. deriving syntactic or semantic representations
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a water turbidity detection method based on a deep regression network, which comprises the following steps of: acquiring water quality images under different turbidities by using an underwater camera, and manually calibrating the acquired images; extracting image features by using a convolutional neural network, and performing feature learning through a plurality of full-connection layers; establishing a mapping relation between the underwater image characteristics and the turbidity value through training a deep regression network; and (3) extracting image characteristics by using a convolutional neural network aiming at a sample image to be detected, and estimating a corresponding turbidity value by using a trained model. According to the invention, a water quality image is acquired through an underwater camera, underwater image characteristics are extracted by using a VggNet network, and then the relationship between the image characteristics and a turbidity value is learned by using a four-layer full-connection network. The method can realize dynamic detection of the turbidity of the water source by using the underwater image.
Description
Technical Field
The invention relates to an intelligent water quality monitoring technology, in particular to a water quality turbidity detection method based on a deep regression network.
Background
Water is an indispensable resource for human beings and other living bodies to live and depend on economic and social development, and water quality monitoring is an important work for realizing reasonable utilization and protection of water resources and sustainable development of economic and social. Turbidity is used as an important index for measuring the water quality condition, and turbidity detection is an important content in water quality monitoring. At present, turbidity is detected mainly by collecting water quality samples of a water source area and adopting a laboratory instrument for detection. The detection mode needs higher manpower and material cost, and is difficult to realize the dynamic detection of turbidity for water with time series characteristics.
CN201710561581.X discloses an underwater calibration target imaging device and an optical image method water turbidity on-line detection method, and the underwater calibration target imaging device can image a calibration target at a known fixed distance. And estimating background light of the underwater scene in the non-calibration target area of the underwater image according to the dark channel prior model. And calculating the water turbidity parameter estimation values of all points on the underwater calibration target according to the distance between all points on the known calibration target and the imaging surface of the same camera in the underwater image calibration target area and on the basis of an underwater optical imaging model. And taking the average value of the water turbidity estimated values obtained according to all points on the calibration target as the detection result of the water turbidity of the current underwater environment. This method does not provide high detection accuracy.
For this case, an image-based depth regression model is proposed for turbidity dynamics monitoring. And acquiring an underwater image of the water source area through an underwater camera, and predicting the turbidity value corresponding to the current monitoring image by using a depth regression model, thereby realizing the dynamic detection of the turbidity of the water source area.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides the water turbidity detection method based on the deep regression network, which can obtain stronger water turbidity representation and improve the detection precision.
The purpose of the invention is realized by the following technical scheme.
A water turbidity detection method based on a deep regression network comprises the following steps:
1) acquiring water quality images under different turbidities by using an underwater camera, and manually calibrating the acquired images;
2) extracting image features by using a convolutional neural network, and performing feature learning through a plurality of full-connection layers;
3) establishing a mapping relation between the underwater image characteristics and the turbidity value through training a deep regression network;
4) and (3) extracting image characteristics by using a convolutional neural network aiming at a sample image to be detected, and estimating a corresponding turbidity value by using a trained model.
Acquiring underwater images of different positions of a water source area according to the cameras arranged at different underwater positions in the step 1) to obtain sample dataWherein x i Representing an image, y i The corresponding turbidity value, n, represents the number of samples.
The image feature extraction method in the step 2) uses a VggNet network, and the dimension of the extracted feature is 4096.
And in the step 2), four layers of fully-connected networks are adopted for image feature learning, and the sizes of the neurons in each layer of network are 1024, 512, 256 and 1 respectively.
The depth regression network in the step 3) adopts a mean square error function, and the calculation formula is as follows:
wherein f (w) represents a model loss function, o i Predicted value of model output, y i And w is the true value, and w is the parameter to be learned by the model.
In the step 3), the gradient of the model parameters is calculated by the deep regression network model by adopting an automatic differentiation technology, and then the model parameters are optimized by adopting a gradient descent method, wherein the formula of the adopted gradient descent method is as follows:
where η is the learning rate.
In the step 4), for the sample image to be detected, firstly extracting features by using a VggNet network, and then predicting the turbidity value of the sample to be detected by using a depth regression network.
Compared with the prior art, the invention has the advantages that: 1. according to the invention, an underwater image is collected, and the VggNet network is adopted to extract image characteristics, so that a stronger water turbidity characterization can be obtained, the subsequent prediction effect is further improved, and the dynamic detection of the water source turbidity is realized.
2. The invention adopts four layers of full-connection network to realize feature learning and turbidity prediction at the same time, thereby improving the detection precision.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The invention is described in detail below with reference to the drawings and specific examples.
As shown in fig. 1, 1) acquiring water quality images under different turbidities by using an underwater camera, and manually calibrating the acquired images;
2) extracting image features by using a convolutional neural network, and performing feature learning through a plurality of full-connection layers;
3) establishing a mapping relation between the underwater image characteristics and the turbidity value through training a deep regression network;
4) and (3) extracting image characteristics by using a convolutional neural network aiming at a sample image to be detected, and estimating a corresponding turbidity value by using a trained model.
Acquiring underwater images of different positions of a water source area according to the cameras arranged at different underwater positions in the step 1) to obtain sample dataWherein x i Representing an image, y i Corresponding turbidity value, n representing the sampleThe number of (2).
The image feature extraction method in the step 2) uses a VggNet network, and the dimension of the extracted feature is 4096.
And in the step 2), four layers of fully-connected networks are adopted for image feature learning, and the sizes of the neurons in each layer of network are 1024, 512, 256 and 1 respectively.
The depth regression network in the step 3) adopts a mean square error function, and the calculation formula is as follows:
wherein f (w) represents a model loss function, o i Predicted value of model output, y i And w is the true value, and w is the parameter to be learned by the model.
In the step 3), the gradient of the model parameters is calculated by the deep regression network model by adopting an automatic differentiation technology, and then the model parameters are optimized by adopting a gradient descent method, wherein the formula of the adopted gradient descent method is as follows:
where η is the learning rate.
In the step 4), for the sample image to be detected, firstly extracting features by using a VggNet network, and then predicting the turbidity value of the sample to be detected by using a depth regression network.
Example 1
In order to verify the prediction model, the method is modeled by using the image data of the water quality of a certain river channel, and the result is as follows:
the water source data comprises 60 images, and each image has a corresponding turbidity value for model training. The size of the image is 200x 200. In the experiment, 40 images are randomly selected as a training set, 11 images are selected as a verification set, and 9 images are selected as a test set. The training set and the verification set are used for model training and parameter adjustment, and the test set is used for verifying the performance of the model. The performance of the proposed method is evaluated by the Mean Square Error (MSE) between the predicted and true values. The results on the test set are shown in the table below.
The mean square error is given in the table as: 0.85. the method for detecting the water turbidity based on the deep regression network has higher precision and can meet the requirements of practical application.
The foregoing is directed to embodiments of the present invention and, more particularly, to a method and apparatus for controlling a power converter in a power converter, including a power converter, a display and a display panel.
Claims (7)
1. A water turbidity detection method based on a deep regression network is characterized by comprising the following steps:
1) acquiring water quality images under different turbidities by using an underwater camera, and manually calibrating the acquired images;
2) extracting image features by using a convolutional neural network, and performing feature learning through a plurality of full-connection layers;
3) establishing a mapping relation between the underwater image characteristics and the turbidity value through training a deep regression network;
4) and (3) extracting image characteristics by using a convolutional neural network aiming at the sample image to be detected, and estimating a turbidity value corresponding to the trained model by using the trained model.
2. The method for detecting the turbidity of water quality based on the deep regression network as claimed in claim 1, wherein in the step 1), underwater images of different positions of a water source are collected according to cameras arranged at different positions underwater to obtain sample dataWherein x i Representing an image, y i The corresponding turbidity value, n, represents the number of samples.
3. The method for detecting the turbidity of water based on the deep regression network as claimed in claim 1, wherein the step 2) image feature extraction method uses a VggNet network, and the extracted feature dimension is 4096.
4. The method for detecting the turbidity of water based on the deep regression network as claimed in claim 1 or 3, wherein the image feature learning of step 2) adopts four layers of fully connected networks, and the size of each layer of network neuron is 1024, 512, 256 and 1.
5. The method for detecting the turbidity of water based on the deep regression network as claimed in claim 1, wherein the deep regression network in the step 3) adopts a mean square error function, and the calculation formula is as follows:
wherein f (w) represents a model loss function, o i Predicted value of model output, y i And w is the true value, and w is the parameter to be learned by the model.
6. The method for detecting the turbidity of water quality based on the deep regression network as claimed in claim 1 or 5, wherein the deep regression network model in the step 3) adopts an automatic differentiation technology to calculate the gradient of the model parameters, and then adopts a gradient descent method to optimize the model parameters, wherein the formula of the adopted gradient descent method is as follows:
where η is the learning rate.
7. The method for detecting the turbidity of water based on the deep regression network as claimed in claim 1, wherein in the step 4), for the image of the sample to be detected, firstly, the VggNet network is used to extract the features, and then the deep regression network is used to predict the turbidity value of the sample to be detected.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210683908.1A CN115063596A (en) | 2022-06-17 | 2022-06-17 | Water quality turbidity detection method based on deep regression network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210683908.1A CN115063596A (en) | 2022-06-17 | 2022-06-17 | Water quality turbidity detection method based on deep regression network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115063596A true CN115063596A (en) | 2022-09-16 |
Family
ID=83201775
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210683908.1A Pending CN115063596A (en) | 2022-06-17 | 2022-06-17 | Water quality turbidity detection method based on deep regression network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115063596A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000218263A (en) * | 1999-02-01 | 2000-08-08 | Meidensha Corp | Water quality controlling method and device therefor |
KR100423115B1 (en) * | 2002-09-19 | 2004-03-18 | (주) 이스텍 | Water Quality Measuring Apparatus and Method Using Image Data |
CN112067517A (en) * | 2020-09-11 | 2020-12-11 | 杭州市地下管道开发有限公司 | Intelligent monitoring method, equipment and system for river and lake water body and readable storage medium |
CN112508106A (en) * | 2020-12-08 | 2021-03-16 | 大连海事大学 | Underwater image classification method based on convolutional neural network |
US20210085165A1 (en) * | 2019-09-24 | 2021-03-25 | Boston Scientific Scimed, Inc. | System, device and method for turbidity analysis |
CN112782097A (en) * | 2020-12-21 | 2021-05-11 | 中国科学院合肥物质科学研究院 | Liquid turbidity measuring device and method based on convolutional neural network |
-
2022
- 2022-06-17 CN CN202210683908.1A patent/CN115063596A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000218263A (en) * | 1999-02-01 | 2000-08-08 | Meidensha Corp | Water quality controlling method and device therefor |
KR100423115B1 (en) * | 2002-09-19 | 2004-03-18 | (주) 이스텍 | Water Quality Measuring Apparatus and Method Using Image Data |
US20210085165A1 (en) * | 2019-09-24 | 2021-03-25 | Boston Scientific Scimed, Inc. | System, device and method for turbidity analysis |
CN112067517A (en) * | 2020-09-11 | 2020-12-11 | 杭州市地下管道开发有限公司 | Intelligent monitoring method, equipment and system for river and lake water body and readable storage medium |
CN112508106A (en) * | 2020-12-08 | 2021-03-16 | 大连海事大学 | Underwater image classification method based on convolutional neural network |
CN112782097A (en) * | 2020-12-21 | 2021-05-11 | 中国科学院合肥物质科学研究院 | Liquid turbidity measuring device and method based on convolutional neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107944396B (en) | Knife switch state identification method based on improved deep learning | |
CN111462076B (en) | Full-slice digital pathological image fuzzy region detection method and system | |
CN110017901B (en) | Infrared spectrum diagnosis method for operating state of electrical equipment in power system | |
CN112818969B (en) | Knowledge distillation-based face pose estimation method and system | |
CN108428231B (en) | Multi-parameter part surface roughness learning method based on random forest | |
CN113436169A (en) | Industrial equipment surface crack detection method and system based on semi-supervised semantic segmentation | |
CN106355195B (en) | System and method for measuring image definition value | |
CN111340771B (en) | Fine particulate matter real-time monitoring method integrating visual information richness and wide-depth joint learning | |
CN111062423B (en) | Point cloud classification method of point cloud graph neural network based on self-adaptive feature fusion | |
CN106897998B (en) | Method and system for predicting information of direct solar radiation intensity | |
CN113780242A (en) | Cross-scene underwater sound target classification method based on model transfer learning | |
CN115690632A (en) | Water environment monitoring method for inland river water body | |
CN115455801A (en) | Sea surface wind speed inversion method and device for HY-2B scanning microwave radiometer based on PSO-DNN | |
CN111223087A (en) | Automatic bridge crack detection method based on generation countermeasure network | |
CN115131747A (en) | Knowledge distillation-based power transmission channel engineering vehicle target detection method and system | |
CN111160100A (en) | Lightweight depth model aerial photography vehicle detection method based on sample generation | |
CN106682604B (en) | Blurred image detection method based on deep learning | |
CN115063596A (en) | Water quality turbidity detection method based on deep regression network | |
CN111881922B (en) | Insulator image recognition method and system based on salient features | |
CN116030292A (en) | Concrete surface roughness detection method based on improved ResNext | |
CN112508946B (en) | Cable tunnel anomaly detection method based on antagonistic neural network | |
CN114972085B (en) | Fine granularity noise estimation method and system based on contrast learning | |
CN116523886A (en) | Water turbidity detection method based on multi-source image depth fusion | |
CN117491357B (en) | Quality monitoring method and system for paint | |
CN116148069B (en) | Bearing tensile strength detection device and detection method |
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 |