CN115100553A - River surface pollution information detection processing method and system based on convolutional neural network - Google Patents
River surface pollution information detection processing method and system based on convolutional neural network Download PDFInfo
- Publication number
- CN115100553A CN115100553A CN202210797762.3A CN202210797762A CN115100553A CN 115100553 A CN115100553 A CN 115100553A CN 202210797762 A CN202210797762 A CN 202210797762A CN 115100553 A CN115100553 A CN 115100553A
- Authority
- CN
- China
- Prior art keywords
- neural network
- river
- colored wastewater
- floating objects
- pollution
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 40
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 22
- 238000003672 processing method Methods 0.000 title claims description 8
- 239000002351 wastewater Substances 0.000 claims abstract description 62
- 238000003062 neural network model Methods 0.000 claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 23
- 238000000034 method Methods 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000012360 testing method Methods 0.000 claims description 13
- 238000012795 verification Methods 0.000 claims description 12
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 10
- 208000034699 Vitreous floaters Diseases 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 230000003321 amplification Effects 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 4
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000003709 image segmentation Methods 0.000 claims description 2
- 239000003344 environmental pollutant Substances 0.000 description 5
- 231100000719 pollutant Toxicity 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000010365 information processing Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000003190 augmentative effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 230000001960 triggered effect Effects 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
- G06V20/10—Terrestrial scenes
- G06V20/17—Terrestrial scenes taken from planes or by drones
-
- 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
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Remote Sensing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Processing (AREA)
Abstract
The invention belongs to the technical field of river pollution detection, and particularly relates to a method and a system for detecting and processing river pollution information based on a convolutional neural network. The method comprises the following steps: s1, collecting the images of the river floating objects and the colored wastewater discharge to form a data set; s2, marking data to obtain coordinate information of the rectangular frame and floating objects and colored wastewater contained in the rectangular frame; s3, amplifying the data set; s4, training the improved Yolov4-tiny neural network model; s5, embedding the improved YOLOv4-tiny neural network model with the optimal performance into an MCU (microprogrammed control Unit), and carrying out video detection on the monitored river or lake; s6, the MCU is installed on the unmanned aerial vehicle, and the pollution degree of the river or lake to be detected is judged according to the detected amount of the floating objects and the colored wastewater. The method can effectively analyze the pollution source and has the characteristics of strong real-time performance and high accuracy.
Description
Technical Field
The invention belongs to the technical field of river pollution detection, and particularly relates to a method and a system for detecting and processing river pollution information based on a convolutional neural network.
Background
At present, China pays great attention to the protection of ecological environment, particularly to the protection of water environment resources. Common river pollution is mainly the discharge of floaters and colored waste water, and man-made river pollution patrols and examines consuming time and power, and unmanned aerial vehicle, unmanned ship river patrol and examine become the leading mode. The prior river pollution detection technology usually adopts river pollutant images for manual interpretation or simple computer target detection, lacks automatic identification and detection of pollutants, and carries out subsequent early warning and information processing and sending, thereby having the problems of low hysteresis and detection accuracy.
Therefore, it is necessary to design a river pollution information detection processing method and system based on a convolutional neural network, which can effectively analyze pollution sources and has the characteristics of strong real-time performance and high accuracy.
For example, chinese patent application No. CN201910059153.6 describes a method for detecting floating objects on water, which comprises the following steps: step 1, collecting data; step 2, enhancing data; data enhancement is performed on data set a. Step 3, marking pictures; marking the floating object area in the data set B by using a rectangular frame, and 4, training a module; the purpose of dividing the data set B into three parts is to be able to select the best-performing, best-generalization-capability weight model. Step 5, detecting the module; and detecting the monitored river or lake video by using the trained weight model. Although the traditional manual-based detection method is replaced, the manpower and material resources are saved, the pollution degree of the river or the lake is judged, the data set can be randomly divided into a training set, a testing set and a verification set, the samples are expanded through a data enhancement method, and the over-fitting problem caused by too few image samples is prevented, the method has the defect that only floaters on the river surface can be detected, and the detection and calculation cannot be carried out on the discharge condition of colored wastewater.
Disclosure of Invention
The invention provides a river pollution information detection processing method and system based on a convolutional neural network, which have the characteristics of strong real-time performance and high accuracy, and aims to overcome the problems that in the prior art, the prior river pollution detection technology usually adopts river pollutant images for manual interpretation or simple computer target detection, automatic identification and detection of pollutants are lacked, and subsequent early warning and information processing and sending are carried out, so that the hysteresis and the detection accuracy are low.
In order to achieve the purpose, the invention adopts the following technical scheme:
the river pollution information detection processing method based on the convolutional neural network comprises the following steps:
s1, collecting the images of the river floating objects and the colored wastewater discharge to form a data set, and classifying the data set into a floating object set and a colored wastewater set;
s2, marking the overwater floating objects and the colored wastewater discharge areas of the two data sets obtained in the step S1 by using a rectangular frame to obtain coordinate information of the rectangular frame and the floating objects and the colored wastewater contained in the rectangular frame;
s3, amplifying the data set and increasing the number of pictures in the data set;
s4, dividing the floating object and the colored waste water into a training set, a verification set and a test set; training the improved Yolov4-tiny neural network model by using a training set; adjusting parameters of the improved YOLOv4-tiny neural network model by using a verification set; selecting an optimal improved YOLOv4-tiny neural network model by using the test set;
s5, embedding the improved YOLOv4-tiny neural network model which is optimal in the step S4 into an MCU, carrying out video detection on the monitored river or lake, and detecting whether floating objects and colored wastewater are discharged on the water surface to be detected in real time;
s6, the MCU is installed on the unmanned aerial vehicle, the unmanned aerial vehicle autonomously cruises along a river channel or a lake by means of the carried GPS, photographs are shot, the photographs are preprocessed, the preprocessed photographs are input into the improved YOLOv4-tiny neural network model, and the pollution degree of the river channel or the lake to be detected is judged according to the detected floating objects and the quantity of colored wastewater.
Preferably, the coordinate information of the rectangular frame in step S2 includes the center point coordinate of the rectangular frame, the width and the height of the rectangular frame.
Preferably, the method for augmenting the data set in step S3 includes flipping, rotating, scaling, cropping, and shifting.
Preferably, the improved YOLOv4-tiny neural network model in step S4 includes a feature extraction network and a prediction network;
the overall network structure of the improved YOLOv4-tiny neural network model has 38 layers, three residual error units are adopted, and the activation function is LeakyReLU; a feature pyramid FPN network is adopted when the effective feature layers are combined;
the feature extraction network includes feature layers with dimensions 13 × 13, 26 × 26, and 52 × 52.
Preferably, the photo preprocessing in step S6 is image cutting, in which a picture of large pixels is cut into a picture of small pixels.
Preferably, the step S6 of judging the pollution level of the river or lake to be detected according to the detected floating objects and the amount of the colored wastewater includes the following steps:
s61, judging that the floating objects are polluted when the number of the detected floating objects exceeds 7; when colored wastewater is detected, the pollution is directly judged.
Preferably, the method further comprises the following steps:
and S7, positioning the current position of the unmanned aerial vehicle through the GPS after the pollution conclusion exists, and sending the current position and the picture to the mobile phone terminal.
The invention also provides a river surface pollution information detection processing system based on the convolutional neural network, which comprises:
the data acquisition module is used for acquiring images of river floating objects and colored wastewater discharge to form a data set, and classifying the data set into a floating object set and a colored wastewater set;
the data marking module is used for marking the overwater floater and the colored wastewater discharge area in the two acquired data sets by using a rectangular frame to obtain coordinate information of the rectangular frame and the floater and the colored wastewater contained in the rectangular frame;
the data amplification module is used for amplifying the data set and increasing the number of pictures in the data set;
the data classification and use module is used for uniformly dividing the floating objects and the colored wastewater into a training set, a verification set and a test set; training the improved Yolov4-tiny neural network model by using a training set; adjusting parameters of the improved YOLOv4-tiny neural network model by using a verification set; selecting an optimal improved YOLOv4-tiny neural network model by using the test set;
the detection module is used for carrying out video detection on the monitored river channel or lake and detecting whether floating objects and colored wastewater are discharged on the water surface to be detected in real time;
and the judging module is used for shooting a picture, preprocessing the picture, inputting the preprocessed picture into the improved YOLOv4-tiny neural network model, and judging the pollution degree of the river or the lake to be detected according to the detected floating objects and the quantity of the colored wastewater.
Preferably, the method further comprises the following steps:
and the information transmission module is used for positioning to the current position through the GPS and sending the current position and the picture to the mobile phone terminal after the pollution conclusion exists.
Compared with the prior art, the invention has the beneficial effects that: (1) the river floating object and colored wastewater are detected in real time based on the YOLOv4-tiny network, the real-time monitoring of the discharge of the water floating object and the colored wastewater can be realized, the discharge conditions of the river floating object and the colored wastewater are detected and calculated, when the colored wastewater and the river floating object are detected to exceed a set value, early warning is triggered, the site conditions are sent to the mobile phone terminal of a worker, the worker can comprehensively know the relevant information of a pollution source in real time, and the pollution source can be effectively analyzed; (2) the method has the advantages of strong real-time performance, high accuracy and easy popularization.
Drawings
FIG. 1 is a flow chart of a river pollution information detection processing method based on a convolutional neural network in the present invention;
FIG. 2 is a schematic structural diagram of an improved Yolov4-tiny neural network model according to the present invention;
fig. 3 is a schematic structural diagram of the unmanned aerial vehicle according to the present invention;
fig. 4 is a schematic block diagram of a system for detecting and processing river pollution information based on a convolutional neural network according to an embodiment of the present invention.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention, the following description will explain the embodiments of the present invention with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
Example (b):
as shown in fig. 1, the invention provides a river pollution information detection processing method based on a convolutional neural network, which comprises the following steps:
s1, collecting the images of the river floating objects and the colored wastewater discharge to form a data set, and classifying the data set into a floating object set and a colored wastewater set;
s2, marking the overwater floating objects and the colored wastewater discharge areas of the two data sets obtained in the step S1 by using a rectangular frame to obtain coordinate information of the rectangular frame and the floating objects and the colored wastewater contained in the rectangular frame; the coordinate information of the rectangular frame comprises the coordinates of the center point of the rectangular frame and the width and the height of the rectangular frame;
s3, amplifying the data set and increasing the number of pictures in the data set; the method for amplifying the data set comprises the steps of turning, rotating, zooming, clipping and shifting;
s4, dividing the floating object and the colored waste water into training set, verifying set and testing set, the proportion is 8: 1: 1; training the improved Yolov4-tiny neural network model by using a training set; adjusting parameters of the improved YOLOv4-tiny neural network model by using a verification set; selecting an optimal improved YOLOv4-tiny neural network model by using the test set;
s5, embedding the improved YOLOv4-tiny neural network model which is optimal in the step S4 into an MCU, carrying out video detection on the monitored river or lake, and detecting whether floating objects and colored wastewater are discharged on the water surface to be detected in real time;
s6, the MCU is installed on the unmanned aerial vehicle, the unmanned aerial vehicle autonomously cruises along a river channel or a lake by means of the carried GPS, photographs are shot, the photographs are preprocessed, the preprocessed photographs are input into the improved YOLOv4-tiny neural network model, and the pollution degree of the river channel or the lake to be detected is judged according to the detected floating objects and the quantity of colored wastewater.
And S7, positioning the current position of the unmanned aerial vehicle through the GPS after the pollution conclusion exists, and sending the current position and the picture to the mobile phone terminal.
As shown in fig. 2, the improved YOLOv4-tiny is composed of two parts, namely, a feature extraction network and a prediction network, the overall network structure has 38 layers, three residual error units are used, the activation function uses LeakyReLU, and a Feature Pyramid (FPN) network is used when effective feature layers are combined. In the aspect of prediction, a detection scale 52 x 52 is added on two feature map bases 13 x 13 and 26 x 26 with different scales on the original main network for prediction to extract shallow information, and a feature layer of 52 x 52 is fused with other two feature layers to form a new feature extraction network, so that shallow features have strong position information, deep features have strong semantic information, and the positioning accuracy is enhanced while the omission ratio of small targets is reduced.
The photo preprocessing in step S6 is image segmentation, which segments a large-pixel picture into small-pixel pictures. Because the pictures that unmanned aerial vehicle took are the picture of extra large pixel, and in the model training, the data set of training all is the data set that probably passes through manual screening, is the clear little pixel picture of target relatively, consequently need cut into the picture of large pixel for the small pixel, is favorable to the detection of small target, carries out subsequent processing again.
The unmanned aerial vehicle shoots a river channel photo, after the photo is divided, the photo is input into the trained neural network model for prediction, three scales of prediction are provided, each scale can predict 3 target frames, for one photo, the photo is initially divided into grids of K x K, and if the center of one object falls on a certain cell, the cell is responsible for detecting the object. The final eigen-map output has a tensor size K × K (3 × 4+1+ C), which includes the 4 centroid coordinates, confidence scores, and object class C needed to determine a target frame. And then setting the border score with the confidence score smaller than the threshold value as 0, finally removing the repeated border box by adopting an NMS (non-maximum suppression) algorithm, and keeping the border box with the maximum score as the final prediction box. The floating material and colored waste water on the picture will be identified.
The step S6 of determining the pollution level of the river or lake to be detected according to the detected amount of the floating objects and the colored wastewater includes the following steps:
s61, judging that the floating objects are polluted when the number of the detected floating objects exceeds 7; when colored wastewater is detected, directly judging the wastewater to be polluted
The invention also provides a river surface pollution information detection processing system based on the convolutional neural network, which comprises:
the data acquisition module is used for acquiring images of river floating objects and colored wastewater discharge to form a data set, and classifying the data set into a floating object set and a colored wastewater set;
the data marking module is used for marking the overwater floater and the colored wastewater discharge area in the two acquired data sets by using a rectangular frame to obtain coordinate information of the rectangular frame and the floater and the colored wastewater contained in the rectangular frame;
the data amplification module is used for amplifying the data set and increasing the number of pictures in the data set;
the data classification and use module is used for uniformly dividing the floating object and the colored wastewater into a training set, a verification set and a test set; training the improved Yolov4-tiny neural network model by using a training set; adjusting parameters of the improved YOLOv4-tiny neural network model by using a verification set; selecting an optimal improved YOLOv4-tiny neural network model by using the test set;
the detection module is used for carrying out video detection on the monitored river channel or lake and detecting whether floating objects and colored wastewater are discharged on the water surface to be detected in real time;
and the judging module is used for shooting a picture, preprocessing the picture, inputting the preprocessed picture into the improved YOLOv4-tiny neural network model, and judging the pollution degree of the river or the lake to be detected according to the detected floating objects and the quantity of the colored wastewater.
And the information transmission module is used for positioning the current position through the GPS and sending the current position and the picture to the mobile phone terminal after the pollution conclusion exists.
Specifically, the adopted equipment is shown in fig. 4 and comprises an unmanned aerial vehicle, a camera, an MCU and a GPS module, wherein an improved YOLOv4-tiny neural network model is trained in advance and then deployed into the MCU to realize real-time recognition of river pollution by embedded equipment, and when the discharge of colored wastewater and floating garbage exceeding a preset value are recognized, the MCU sends related information to a mobile phone terminal of a worker through a 4G module.
The data acquisition module specifically adopts an unmanned aerial vehicle camera; the data marking module, the data amplification module, the data classification and use module, the detection module and the judgment module are all included in the MCU; the information transmission module specifically adopts a 4G module.
The MCU of the invention specifically adopts the Jetson Nano A02. The English viada Jetson Nano A02 is used as the embedded device, a four-core 64-bit ARM CPU and a 128-core are integrated into an NVIDIA GPU, the computing performance of 472GFLOPS can be provided, and the computing power supports the operation of YOLOv4-tiny for detection and identification.
The GPS module adopts a Chinese micro AT6558R positioning chip, can receive satellite signals by a 99 channel and has low power consumption; the high-sensitivity G-MOUSE can be used for quickly and accurately positioning in places with weak signals such as cities, canyons, under high buildings and the like and any position in the automobile. Make the GPS module can extensively be used for on-vehicle control, bus stop report, on-vehicle navigation, ship navigation, notebook navigation etc. product, consequently this kind of chip can satisfy unmanned aerial vehicle's location demand.
In addition, the appearance structure of the unmanned aerial vehicle is shown in fig. 3.
The invention improves the accuracy of pollutant detection, accurately classifies the pollution and quickly and accurately finds the discharge source of the colored wastewater. The invention can realize the detection function only by carrying the embedded equipment of the lightweight convolutional neural network on the price, and has the advantages of high cost performance and strong practicability. The invention accurately positions through the GPS, and is convenient for finding the position of the pollution source. According to the invention, the 4G module is used for sending the relevant information to the mobile phone terminal, so that the field situation can be rapidly known. The unmanned aerial vehicle automatic cruise control system adopts the unmanned aerial vehicle to automatically cruise, reduces manual cruise, and is time-saving and labor-saving. The invention can improve the protection of the ecological environment, in particular to the protection of the water environment.
The foregoing has outlined rather broadly the preferred embodiments and principles of the present invention and it will be appreciated that those skilled in the art may devise variations of the present invention that are within the spirit and scope of the appended claims.
Claims (9)
1. The river surface pollution information detection processing method based on the convolutional neural network is characterized by comprising the following steps:
s1, collecting the images of the river floating objects and the colored wastewater discharge to form a data set, and classifying the data set into a floating object set and a colored wastewater set;
s2, marking the overwater floating objects and the colored wastewater discharge areas of the two data sets obtained in the step S1 by using a rectangular frame to obtain coordinate information of the rectangular frame and the floating objects and the colored wastewater contained in the rectangular frame;
s3, amplifying the data set and increasing the number of pictures in the data set;
s4, dividing the floating objects and the colored wastewater into a training set, a verification set and a test set; training the improved Yolov4-tiny neural network model by using a training set; adjusting parameters of the improved YOLOv4-tiny neural network model by using a verification set; selecting an optimal improved YOLOv4-tiny neural network model by using the test set;
s5, embedding the improved YOLOv4-tiny neural network model which is optimal in the step S4 into an MCU, carrying out video detection on the monitored river or lake, and detecting whether floating objects and colored wastewater are discharged on the water surface to be detected in real time;
s6, the MCU is installed on the unmanned aerial vehicle, the unmanned aerial vehicle autonomously cruises along a river or lake by means of the carried GPS, takes a picture, then preprocesses the picture, inputs the preprocessed picture into the improved YOLOv4-tiny neural network model, and judges the pollution degree of the river or lake to be detected according to the quantity of detected floaters and colored wastewater.
2. The method for detecting and processing river pollution information based on the convolutional neural network as claimed in claim 1, wherein the coordinate information of the rectangular frame in step S2 includes the coordinates of the center point of the rectangular frame, the width and the height of the rectangular frame.
3. The method for detecting and processing the river pollution information based on the convolutional neural network of claim 1, wherein the method for amplifying the data set in step S3 comprises flipping, rotating, scaling, cropping and shifting.
4. The method for detecting and processing river pollution information based on convolutional neural network of claim 3, wherein the improved Yolov4-tiny neural network model in step S4 comprises a feature extraction network and a prediction network;
the overall network structure of the improved YOLOv4-tiny neural network model has 38 layers, three residual error units are adopted, and the activation function is LeakyReLU; a feature pyramid FPN network is adopted when the effective feature layers are combined;
the feature extraction network includes layers of features with dimensions 13 × 13, 26 × 26, and 52 × 52.
5. The method for detecting and processing river pollution information based on the convolutional neural network as claimed in claim 5, wherein the step of preprocessing the picture in step S6 is image segmentation, and a picture with large pixels is segmented into a picture with small pixels.
6. The method for detecting and processing river pollution information based on convolutional neural network as claimed in claim 1, wherein said determining the pollution level of the river or lake to be detected according to the amount of the detected floating objects and colored wastewater in step S6 comprises the following steps:
s61, judging that the floating objects are polluted when the number of the detected floating objects exceeds 7; when colored wastewater is detected, directly judging the wastewater to be polluted.
7. The method for detecting and processing river pollution information based on the convolutional neural network as claimed in claim 1, further comprising the following steps:
and S7, positioning the current position of the unmanned aerial vehicle through the GPS after the pollution conclusion exists, and sending the current position and the picture to the mobile phone terminal.
8. River surface pollution information detection processing system based on convolutional neural network, its characterized in that includes:
the data acquisition module is used for acquiring images of river floating objects and colored wastewater discharge to form a data set, and classifying the data set into a floating object set and a colored wastewater set;
the data marking module is used for marking the overwater floater and the colored wastewater discharge area in the two acquired data sets by using a rectangular frame to obtain coordinate information of the rectangular frame and the floater and the colored wastewater contained in the rectangular frame;
the data amplification module is used for amplifying the data set and increasing the number of pictures in the data set;
the data classification and use module is used for uniformly dividing the floating objects and the colored wastewater into a training set, a verification set and a test set; training the improved YOLOv4-tiny neural network model by using a training set; adjusting parameters of the improved YOLOv4-tiny neural network model by using a verification set; selecting an optimal improved YOLOv4-tiny neural network model by using the test set;
the detection module is used for carrying out video detection on the monitored river channel or lake and detecting whether floating objects and colored wastewater are discharged on the water surface to be detected in real time;
and the judging module is used for shooting a picture, preprocessing the picture, inputting the preprocessed picture into the improved YOLOv4-tiny neural network model, and judging the pollution degree of the river or the lake to be detected according to the detected floating objects and the quantity of the colored wastewater.
9. The convolutional neural network-based river pollution information detection processing system as claimed in claim 8, further comprising:
and the information transmission module is used for positioning to the current position through the GPS and sending the current position and the picture to the mobile phone terminal after the pollution conclusion exists.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210797762.3A CN115100553A (en) | 2022-07-06 | 2022-07-06 | River surface pollution information detection processing method and system based on convolutional neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210797762.3A CN115100553A (en) | 2022-07-06 | 2022-07-06 | River surface pollution information detection processing method and system based on convolutional neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115100553A true CN115100553A (en) | 2022-09-23 |
Family
ID=83295954
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210797762.3A Pending CN115100553A (en) | 2022-07-06 | 2022-07-06 | River surface pollution information detection processing method and system based on convolutional neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115100553A (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109033934A (en) * | 2018-05-25 | 2018-12-18 | 江南大学 | A kind of floating on water surface object detecting method based on YOLOv2 network |
CN111062316A (en) * | 2019-12-16 | 2020-04-24 | 成都之维安科技股份有限公司 | Pollution source wastewater discharge real-time video analysis system based on deep learning technology |
CN112488020A (en) * | 2020-12-10 | 2021-03-12 | 西安交通大学 | Water environment pollution condition detection and evaluation device based on unmanned aerial vehicle aerial photography data |
WO2021142902A1 (en) * | 2020-01-17 | 2021-07-22 | 五邑大学 | Danet-based unmanned aerial vehicle coastline floating garbage inspection system |
-
2022
- 2022-07-06 CN CN202210797762.3A patent/CN115100553A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109033934A (en) * | 2018-05-25 | 2018-12-18 | 江南大学 | A kind of floating on water surface object detecting method based on YOLOv2 network |
CN111062316A (en) * | 2019-12-16 | 2020-04-24 | 成都之维安科技股份有限公司 | Pollution source wastewater discharge real-time video analysis system based on deep learning technology |
WO2021142902A1 (en) * | 2020-01-17 | 2021-07-22 | 五邑大学 | Danet-based unmanned aerial vehicle coastline floating garbage inspection system |
CN112488020A (en) * | 2020-12-10 | 2021-03-12 | 西安交通大学 | Water environment pollution condition detection and evaluation device based on unmanned aerial vehicle aerial photography data |
Non-Patent Citations (1)
Title |
---|
唐小敏;舒远仲;刘文祥;刘金梅;: "基于SSD深度网络的河道漂浮物检测技术研究", 计算机技术与发展, no. 09, 10 September 2020 (2020-09-10) * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Wu et al. | Road pothole extraction and safety evaluation by integration of point cloud and images derived from mobile mapping sensors | |
CN108460341B (en) | Optical remote sensing image target detection method based on integrated depth convolution network | |
CN109711264B (en) | Method and device for detecting occupation of bus lane | |
CN110208880B (en) | Sea fog detection method based on deep learning and satellite remote sensing technology | |
CN111626170B (en) | Image recognition method for railway side slope falling stone intrusion detection | |
CN114998852A (en) | Intelligent detection method for road pavement diseases based on deep learning | |
CN111126183A (en) | Method for detecting damage of building after earthquake based on near-ground image data | |
CN106934374A (en) | The recognition methods of traffic signboard and system in a kind of haze scene | |
CN112666553B (en) | Road ponding identification method and equipment based on millimeter wave radar | |
CN112926583B (en) | License plate recognition method and license plate recognition system | |
Amrutha et al. | Real-time litter detection system for moving vehicles using YOLO | |
CN111753610A (en) | Weather identification method and device | |
Li et al. | Automated classification and detection of multiple pavement distress images based on deep learning | |
Kharel et al. | Potholes detection using deep learning and area estimation using image processing | |
CN116228756B (en) | Method and system for detecting bad points of camera in automatic driving | |
CN111597939B (en) | High-speed rail line nest defect detection method based on deep learning | |
CN113311507A (en) | Typhoon path identification method and device | |
KR102259981B1 (en) | Visual range measurement and alarm system based on video analysis and method thereof | |
CN108694415A (en) | Image characteristic extracting method, device and water source image classification method, device | |
CN115100553A (en) | River surface pollution information detection processing method and system based on convolutional neural network | |
Anagnostopoulos et al. | Using sliding concentric windows for license plate segmentation and processing | |
CN111339823A (en) | Threshing and sunning ground detection method based on machine vision and back projection algorithm | |
CN110765900A (en) | DSSD-based automatic illegal building detection method and system | |
Pan et al. | Identifying Vehicles Dynamically on Freeway CCTV Images through the YOLO Deep Learning Model. | |
CN112241691B (en) | Channel ice condition intelligent identification method based on unmanned aerial vehicle inspection and image characteristics |
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 |