CN113326783A - Edge early warning method for water conservancy industry - Google Patents
Edge early warning method for water conservancy industry Download PDFInfo
- Publication number
- CN113326783A CN113326783A CN202110607203.7A CN202110607203A CN113326783A CN 113326783 A CN113326783 A CN 113326783A CN 202110607203 A CN202110607203 A CN 202110607203A CN 113326783 A CN113326783 A CN 113326783A
- Authority
- CN
- China
- Prior art keywords
- water conservancy
- warning method
- conservancy industry
- edge
- target
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the field of intelligent hydrology, and particularly provides an edge early warning method for water conservancy industry, which comprises the following steps: s1, capturing images of the place to be early-warned and carrying out digital processing; s2, training the sample by using a convolutional neural network to form an effective algorithm; and S3, close-range shooting measurement processing, determining the shape and motion state of the target and acquiring instantaneous information of the target. Compared with the prior art, the invention improves the efficiency of image video analysis, effectively reduces the phenomena of false alarm and missing report, reduces the amount of useless data and improves the response speed.
Description
Technical Field
The invention relates to the field of intelligent hydrology, and particularly provides an edge early warning method for the water conservancy industry.
Background
At present, the wading event is discovered and then processed mainly by three modes, namely manual inspection, mass report and traditional monitoring. Wherein the content of the first and second substances,
the manual inspection is as follows: the patrolman records the patrolling item through a mode of manually transcribing data. And reporting the abnormal conditions after the patrol is finished, and summarizing the abnormal conditions into a patrol report. The method is simple and easy to learn, the requirement on operators is not high, but the patrol route needs to be kept in mind, and when the copied data is wrong, the modification is inconvenient. Meanwhile, manual inspection is easy to cause errors and false alarm. Meanwhile, 24-hour uninterrupted inspection cannot be achieved through manual inspection, the manual inspection cost is huge, the inspection safety coefficient is low, and the efficiency is low.
The report by the masses is as follows: people can report illegal wading and illegal behaviors by means of letters, calls, websites and the like, and necessary evidence materials such as names, finding time, specific positions (preferably positioning), photos, video and the like of reported objects, real names and contact ways of the reporters are required to be provided when the reporters report. And (4) checking the reported condition by related departments, and then modifying the event according to responsibility and division. This method is time-consuming and completely dependent on manpower, and is inefficient.
The traditional monitoring is as follows: the person on duty needs to stare at the monitoring screen for a long time, fatigue is easy to generate, attention is reduced, important picture information is missed, and warning is delayed. The number of the water affair image monitoring points is large, and monitors are difficult to be configured for the monitoring cameras according to the ratio of 1:1, so that the water affair image monitoring points are mostly displayed on the monitors in a split screen arrangement mode, and the abnormal phenomenon is likely to be missed by negligence. Because the traditional video does not have an intelligent analysis function, the video data is simply marked with a time tag and stored. When some section of abnormality needs to be investigated, a large amount of time is needed for searching and analyzing, and the working efficiency is very low.
Therefore, the problems of low intelligent means, dependence on manpower and low efficiency exist in the prior art for wading events.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides an edge early warning method with strong practicability in the water conservancy industry.
The technical scheme adopted by the invention for solving the technical problems is as follows:
an edge early warning method for water conservancy industry comprises the following steps:
s1, capturing images of the place to be early-warned and carrying out digital processing;
s2, training the sample by using a convolutional neural network to form an effective algorithm;
and S3, close-range shooting measurement processing, determining the shape and motion state of the target and acquiring instantaneous information of the target.
Further, in step S1, an image is captured by the image pickup device and then transferred to the processing unit for digital processing.
Further, after the digital processing, the image is subjected to preprocessing operation, and the type, size, shape and color of the sample are determined according to the basic information of the pixel.
Preferably, the preprocessing operation includes performing type transformation, de-drying filtering, overlay smoothing and binarization on the image, and performing discrimination of sample type, size, shape and color according to pixel distribution, brightness and color.
Further, in step S2, the convolutional neural network is used to train the sample features, input the raw data, and strip the high-level semantic concepts from the raw data through convolution, activation function and pooling.
Further, after the high-level semantic concept is stripped, the error value between the real value and the output value is calculated through an error function, feedback is carried out layer by layer in the reverse direction, parameters of each layer are updated, and finally the model is converged through feedforward operation and feedback operation to complete training.
Preferably, when the convolutional neural network is used for training the sample characteristics, a sample with less noise is selected for training the convolutional neural network.
Further, in step S3, the target image is captured at a close distance, processed to determine the shape and motion state of the target, so as to obtain a lot of geometric and physical information of the target instantly, and the spatial position of multiple points at a certain instant is determined, so that the photo data can be compared at any time.
Compared with the prior art, the edge early warning method for the water conservancy industry has the following outstanding beneficial effects:
aiming at the requirements of water conservancy industry management and supervision of various events, the image acquisition unit acquires scene information in real time through image video, analyzes the scene image in real time through machine vision analysis and in combination with an image processing algorithm after related deep learning, fully plays a role of a video technology in water conservancy informatization, improves the efficiency of image video analysis, effectively reduces the phenomena of false alarm and missing report, reduces the amount of useless data and improves the response speed.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flow diagram of an edge early warning method in the water conservancy industry.
Detailed Description
The present invention will be described in further detail with reference to specific embodiments in order to better understand the technical solutions of the present invention. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A preferred embodiment is given below:
as shown in fig. 1, an edge early warning method for water conservancy industry in this embodiment includes the following steps:
s1, capturing images of a place needing early warning and carrying out digital processing:
the image is captured by an image capturing device, then the image is transmitted to a processing unit, a series of image preprocessing operations such as type conversion, denoising filtering, overlaying smoothing and binarization are carried out on the image through digital processing, and the type, size, shape, color and the like of a sample are judged according to information such as pixel distribution, brightness, color and the like.
S2, training the sample by using the convolutional neural network to form an effective algorithm:
the method is characterized in that a convolutional neural network is utilized to train various sample characteristics, the algorithm implementation process is to input original data, and high-level semantic concepts in the original data are stripped through layer-by-layer operations such as convolution, activation functions, pooling and the like.
And calculating an error value between the real value and the output value through an error function, reversely feeding back layer by layer, updating parameters of each layer, and finally converging the model through feedforward operation and feedback operation to fulfill the aim of training.
Meanwhile, the performance of the convolutional neural network is closely related to the provided training samples, and the good training samples need to meet the requirements of both the number of the samples and the quality of the samples.
Therefore, it is crucial to collect a large number of high quality samples, directly affecting the accuracy of the final observations. Considering that much interference is removed during segmentation, samples with little noise are selected for training the convolutional neural network.
S3, close-range camera shooting measurement processing, determining the shape and motion state of the target and acquiring the instantaneous information of the target:
the image of the target is shot in a close range, the appearance and the motion state of the target are determined after processing, a large amount of geometric and physical information of the target to be detected can be obtained instantly, the spatial position of a plurality of points at a certain instant is determined, and the photo data can be compared at any time.
The method is mainly applied to the following scenes:
and (3) intrusion detection of water source areas/spillways/reservoirs and the like:
the perimeter precaution identification is based on various methods (such as interframe difference, background difference and the like) for extracting moving targets in video images, pedestrian and vehicle images are extracted, a water conservancy general object training image set is established, a perimeter intelligent analysis and calculation capacity and a human and vehicle object identification library (an object detection model based on a neural network) are established according to the appearance contour characteristics of pedestrians and vehicles, and in a streaming medium of video monitoring and shooting, people swimming (children), river fishing, flood discharge canal (ditch) personnel intrusion, mountain torrent ditch driving vehicles/pedestrians and the like in a common reservoir are identified, early warning is carried out, and warning information is given to business personnel.
River floating object identification:
at present, the river-growing patrolling mainly depends on three modes of manual patrolling, mass reporting and traditional monitoring and then processing, the intelligent means is low, the manpower is relied on, and the efficiency is not high. The river floater is analyzed by video analysis, so that the danger of river patrol personnel is reduced, and the patrol efficiency is improved.
By means of computer vision and image processing, river floating objects are automatically found, automatic early warning is achieved, found floating object events are pushed to government affair WeChat of responsible patrolmen, and the aim of rapid disposal is achieved.
And (3) land line encroachment:
the method is characterized by comprising the steps of yellow river high-quality development and Yangtze river high-quality development, wherein ecological protection of the yellow river and the Yangtze river is developed, and all-weather identification of river channel violation, illegal construction, ecological damage and the like is realized. The device can be used for assisting the supervision of the high-quality development of the yellow river and the high-quality development of the Yangtze river.
An encroaching water area shoreline recognition model is built, an encroaching river behavior is recognized through a water area shoreline of a defined target area, once an illegal behavior is found, an alarm is triggered immediately, and meanwhile, the cloud supports recording of a field live situation to serve as a accountability basis of the illegal behavior.
And (4) supervision of a sewage draining outlet:
today, the discharge of industrial wastewater and domestic sewage is a major cause of water quality pollution. Some sewage draining outlets which do not meet the requirements of the laws and regulations are usually concealed, and the sewage draining period is unstable, such as the phenomenon of avoiding the drainage in daytime, which increases the difficulty of the inspection and the cleaning. In order to meet the requirement of protecting the water environment, the environment protection department needs to check the illegal and unreasonable sewage draining outlets, and then support data is provided for the renovation work.
Patrol the steal drain through the unmanned aerial vehicle based on the intelligent detection method of the peripheral drain in the water area of the thermal infrared image, utilize monitoring facilities (video + current meter) to check the current drain, supervise the multirow, utilize the water quality analyzer to monitor the sewage content, utilize unmanned ship to sample, preserve water quality fast simultaneously. And the pollution discharge supervision is realized in all directions.
The above embodiments are only specific cases of the present invention, and the protection scope of the present invention includes but is not limited to the above embodiments, and any suitable changes or substitutions according to the claims of the edge warning method of the water conservancy industry and any person of ordinary skill in the art should fall within the protection scope of the present invention.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (8)
1. An edge early warning method for water conservancy industry is characterized by comprising the following steps:
s1, capturing images of the place to be early-warned and carrying out digital processing;
s2, training the sample by using a convolutional neural network to form an effective algorithm;
and S3, close-range shooting measurement processing, determining the shape and motion state of the target and acquiring instantaneous information of the target.
2. The edge warning method for the water conservancy industry according to claim 1, wherein in step S1, the image is captured by an image capturing device and then transmitted to a processing unit for digital processing.
3. The edge warning method for the water conservancy industry according to claim 2, wherein after the digital processing, the image is preprocessed, and the type, size, shape and color of the sample are determined according to the basic information of the pixels.
4. The edge early warning method for the water conservancy industry according to claim 3, wherein the preprocessing operation comprises type transformation, drying filtering, overlaying smoothing and binarization on the image, and distinguishing the type, size, shape and color of the sample according to pixel distribution, brightness and color.
5. The edge early warning method for the water conservancy industry according to claim 4, wherein in step S2, a convolutional neural network is used for training sample characteristics, raw data is input, and high-level semantic concepts in the raw data are stripped through convolution, an activation function and pooling operations.
6. The edge early warning method for the water conservancy industry according to claim 5, wherein after the high-level semantic concepts are stripped, an error value between a real value and an output value is calculated through an error function, feedback is performed layer by layer in a reverse direction, parameters of each layer are updated, and finally a model is converged through feed-forward operation and feedback operation to complete training.
7. The edge early warning method for the water conservancy industry according to claim 6, wherein when the convolutional neural network is used for training the sample characteristics, a sample with low noise is selected for training the convolutional neural network.
8. The edge warning method for water conservancy industry according to claim 7, wherein in step S3, the target image is captured in a close range, processed to determine the shape and motion state of the target, so as to obtain a lot of geometric and physical information of the target instantly, and the spatial positions of multiple points at a certain instant are determined, so that the shot data can be compared at any time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110607203.7A CN113326783A (en) | 2021-06-01 | 2021-06-01 | Edge early warning method for water conservancy industry |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110607203.7A CN113326783A (en) | 2021-06-01 | 2021-06-01 | Edge early warning method for water conservancy industry |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113326783A true CN113326783A (en) | 2021-08-31 |
Family
ID=77422974
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110607203.7A Pending CN113326783A (en) | 2021-06-01 | 2021-06-01 | Edge early warning method for water conservancy industry |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113326783A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116186121A (en) * | 2023-04-24 | 2023-05-30 | 四川边缘算力科技有限公司 | Early warning method, system and storage medium based on edge algorithm |
| CN116306223A (en) * | 2023-01-09 | 2023-06-23 | 浪潮智慧科技有限公司 | Hydraulic engineering monitoring method, equipment and medium |
| CN116310760A (en) * | 2023-04-06 | 2023-06-23 | 河南禹宏实业有限公司 | Intelligent water conservancy monitoring system based on machine vision |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110148283A (en) * | 2019-05-16 | 2019-08-20 | 安徽天帆智能科技有限责任公司 | It is a kind of to fall water monitoring system in real time based on convolutional neural networks |
| CN110210323A (en) * | 2019-05-09 | 2019-09-06 | 浙江大学 | A kind of drowned behavior online recognition method based on machine vision |
-
2021
- 2021-06-01 CN CN202110607203.7A patent/CN113326783A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110210323A (en) * | 2019-05-09 | 2019-09-06 | 浙江大学 | A kind of drowned behavior online recognition method based on machine vision |
| CN110148283A (en) * | 2019-05-16 | 2019-08-20 | 安徽天帆智能科技有限责任公司 | It is a kind of to fall water monitoring system in real time based on convolutional neural networks |
Non-Patent Citations (3)
| Title |
|---|
| 李森浩: "基于特征融合的小型水域漂浮物识别方法研究与实现", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅰ辑》 * |
| 王战友 等: "边缘计算终端在河湖漂浮物智能识别上的应用", 《中国水利学会2020学术年会论文集》 * |
| 雷佳明 等: "人工智能图像识别在水库智能监控方面的应用", 《中国水利学会2019学术年会论文集》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116306223A (en) * | 2023-01-09 | 2023-06-23 | 浪潮智慧科技有限公司 | Hydraulic engineering monitoring method, equipment and medium |
| CN116306223B (en) * | 2023-01-09 | 2023-09-22 | 浪潮智慧科技有限公司 | Hydraulic engineering monitoring method, equipment and medium |
| CN116310760A (en) * | 2023-04-06 | 2023-06-23 | 河南禹宏实业有限公司 | Intelligent water conservancy monitoring system based on machine vision |
| CN116310760B (en) * | 2023-04-06 | 2023-09-22 | 河南禹宏实业有限公司 | Intelligent water conservancy monitoring system based on machine vision |
| CN116186121A (en) * | 2023-04-24 | 2023-05-30 | 四川边缘算力科技有限公司 | Early warning method, system and storage medium based on edge algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113326783A (en) | Edge early warning method for water conservancy industry | |
| CN110232380B (en) | Fire night scene restoration method based on Mask R-CNN neural network | |
| Carletti et al. | An intelligent flying system for automatic detection of faults in photovoltaic plants | |
| CN110084165B (en) | Intelligent identification and early warning method for abnormal events in open scene of power field based on edge calculation | |
| CN112633231B (en) | Fire disaster identification method and device | |
| TW202013252A (en) | License plate recognition system and license plate recognition method | |
| CN103106766A (en) | Forest fire identification method and forest fire identification system | |
| CN110659546B (en) | Illegal booth detection method and device | |
| CN105915840A (en) | Factory smoke emission automatic monitoring method based on video signal | |
| CN114170569A (en) | Method, system, storage medium and equipment for monitoring road surface abnormal condition | |
| US11580659B2 (en) | Method for size estimation by image recognition of specific target using given scale | |
| Tao et al. | Smoky vehicle detection based on range filtering on three orthogonal planes and motion orientation histogram | |
| CN115497015A (en) | River floating pollutant identification method based on convolutional neural network | |
| JP2007065878A (en) | Moving object counting method on water surface or underwater, and counting device and counting program | |
| Lam et al. | Real-time traffic status detection from on-line images using generic object detection system with deep learning | |
| CN111783700A (en) | Automatic recognition early warning method and system for road foreign matters | |
| CN102244769A (en) | Object and key person monitoring system and method thereof | |
| KR102040562B1 (en) | Method to estimate visibility distance using image information | |
| CN113221760A (en) | Expressway motorcycle detection method | |
| CN116071711B (en) | Traffic jam condition detection method and device | |
| CN115661475A (en) | Image foreign matter identification method, device, equipment and storage medium | |
| CN113627388A (en) | Method and device for determining bare soil coverage condition | |
| CN114266837A (en) | Road sign detection method, device, equipment and computer storage medium | |
| CN111382697A (en) | Image data processing method and first electronic equipment | |
| Lin et al. | A Garbage Recognition Algorithm Based on YoloX Using in Flood Control System |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210831 |