CN110223341A - A kind of Intelligent water level monitoring method based on image recognition - Google Patents
A kind of Intelligent water level monitoring method based on image recognition Download PDFInfo
- Publication number
- CN110223341A CN110223341A CN201910517818.3A CN201910517818A CN110223341A CN 110223341 A CN110223341 A CN 110223341A CN 201910517818 A CN201910517818 A CN 201910517818A CN 110223341 A CN110223341 A CN 110223341A
- Authority
- CN
- China
- Prior art keywords
- water
- water gauge
- image
- gauge
- height
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 229
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000012544 monitoring process Methods 0.000 title claims abstract description 25
- 238000001514 detection method Methods 0.000 claims abstract description 15
- 238000013528 artificial neural network Methods 0.000 claims abstract description 4
- 239000002352 surface water Substances 0.000 claims abstract description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 20
- 238000000605 extraction Methods 0.000 claims description 14
- 230000000306 recurrent effect Effects 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 3
- 229910017435 S2 In Inorganic materials 0.000 claims 1
- 238000005259 measurement Methods 0.000 abstract description 5
- 238000013135 deep learning Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000005286 illumination Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 230000004438 eyesight Effects 0.000 description 2
- 238000007667 floating Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 101100524346 Xenopus laevis req-a gene Proteins 0.000 description 1
- 101100524347 Xenopus laevis req-b gene Proteins 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000013095 identification testing Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000003595 mist Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000004297 night vision Effects 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 238000009428 plumbing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000005413 snowmelt Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01F—MEASURING VOLUME, VOLUME FLOW, MASS FLOW OR LIQUID LEVEL; METERING BY VOLUME
- G01F23/00—Indicating or measuring liquid level or level of fluent solid material, e.g. indicating in terms of volume or indicating by means of an alarm
- G01F23/04—Indicating or measuring liquid level or level of fluent solid material, e.g. indicating in terms of volume or indicating by means of an alarm by dip members, e.g. dip-sticks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Fluid Mechanics (AREA)
- Geometry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The Intelligent water level monitoring method based on image recognition that the invention proposes a kind of, comprising: receive the image data sent by picture pick-up device in tested region；Detection identification is carried out to the above water gauge part of the water surface in described image data using deep neural network, the water gauge image that algorithm is partitioned into the water surface or more is cut using figure on this basis；The water gauge height of the water surface or more is calculated according to the water gauge image being partitioned into, and combines the whole height of water gauge, obtains water surface water gauge height below；The height of the elevation and the following water gauge of the water surface that obtain real site water gauge bottom can calculate water level value.The present invention is not necessarily to artificial on-site land survey water level, realizes that image obtains by remote mode, is analyzed and processed locally, calculates water level value, significantly reduce manual work cost, also improve the accuracy of measurement.
Description
Technical field
The present invention relates to water level monitoring technical field, in particular to a kind of Intelligent water level monitoring side based on image recognition
Method.
Background technique
Water-level observation refers to the measurement on the spot to the water level of rivers, lake and underground water etc..Water level prediction and human society
Life and the relations of production are close.Planning, design, construction and the management of hydraulic engineering need water level prediction.Bridge, harbour, navigation channel,
The engineering constructions such as plumbing also need water level prediction.In flood-control and drought relief, water level prediction is even more important, it is hydrologic forecast and hydrology feelings
The foundation of report.Water level prediction is all important in the research of stage discharge relation and in the analysis of river load, ice condition etc.
Basic document.
Usually measured using water gauge.Water gauge is traditional effective direct observation device.When actual measurement, the reading on water gauge adds
Water gauge zero point elevation is up to water level.Observation time and observation frequency will adapt to the process of SEA LEVEL VARIATION in one day, meet the hydrology
The requirement of forecast and hydrographic message.Under normal circumstances, day surveys 1~2 time.There is flood, icing, floating ice, generate ice dam and have ice
When snowmelt feeder, increases observation frequency, the result measured is enable completely to reflect the process of SEA LEVEL VARIATION.
Existing mode is this mode not only labor intensive by artificial on-site land survey, but also is deposited to personal safety
It is threatening, and efficiency is lower.
Summary of the invention
The purpose of the present invention aims to solve at least one of described technological deficiency.
For this purpose, it is an object of the invention to propose a kind of Intelligent water level monitoring method based on image recognition.
To achieve the goals above, the embodiment of the present invention provides a kind of Intelligent water level monitoring side based on image recognition
Method includes the following steps:
Step S1 receives the image data sent by picture pick-up device in tested region, wherein wrap in described image data
Include water gauge image；
Step S2 carries out detection identification to the water gauge in described image data using deep neural network, on this basis
The water gauge image that algorithm is partitioned into the water surface or more is cut using figure；
Step S3 calculates the water gauge height of the water surface or more according to the water gauge image being partitioned into, and combines the entirety of water gauge
Height, obtain water surface water gauge height below；
Step S4, the height for obtaining the following water gauge of the water surface obtained in the elevation and step S3 of real site water gauge bottom can
To calculate water level value.
Further, in the step S2, the water gauge in described image data is carried out using Faster R-CNN algorithm
Detection identification.
Further, using the feature extraction network of Faster R-CNN algorithm, water is extracted from the image data received
The feature of ruler image generates characteristic pattern；
Using RPN network as candidate region network, handled on the characteristic pattern, output have a variety of scales and
The rectangular target candidate region of the ratio of width to height；
It is defeated according to the feature in candidate region by characteristic pattern and the object candidate area of generation input classification Recurrent networks
The classification and water gauge bounding box for the water gauge candidate region being born.
Further, it according to preset area-of-interest, is sat using the top left co-ordinate of the area-of-interest and the lower right corner
It is denoted as cutting the input coordinate of algorithm for figure, the water gauge figure that algorithm is partitioned into the water surface or more from the water gauge bounding box is cut using figure
Picture, wherein the area-of-interest is the default region for confining water gauge.
Further, read in the top left co-ordinate slave site water gauge data table of the area-of-interest, bottom right angular coordinate with
The lower-left angular coordinate of the water gauge bounding box is consistent.
Further, the feature extraction network of the Faster R-CNN algorithm uses following one kind: ZF network, VGG16 net
Network and AlexNet network.
Further, in the step S3, the water gauge that the water gauge image that the basis is partitioned into calculates the water surface or more is high
Degree includes the following steps: according to the water gauge image for being partitioned into the water surface or more, using the pixel number of the above water gauge of the water surface divided by unit
Pixel number in length obtains the above water gauge height of the water surface, then subtracts the above water gauge height of the water surface by the whole height of water gauge and be
The following water gauge height of the water surface.
Further, pixel number is calculated by preset website water gauge data table in the unit length.
Intelligent water level monitoring method according to an embodiment of the present invention based on image recognition, the water based on intelligent image identification
Level monitoring system, using relevant computer vision technique and machine learning (including deep learning) algorithm, exploitation based on figure
As the intelligent water gauge reading technology of (video), it can receive what the picture pick-up device teletransmission that reservoir, tunnel etc. are installed came
Image or video containing water gauge, are combined by image processing techniques, deep learning and conventional method, to the water gauge in image
Detected and identified, demarcated the above water gauge part of the water surface, calculate the above water gauge Partial Height of the water surface etc., and then obtain water level number
According to.Realization is automatically saved, shown or is returned request end to calculated waterlevel data.If water level value is more than warning value, can
To carry out water level early warning.The present invention is not necessarily to artificial on-site land survey water level, realizes that image obtains by remote mode, in local progress
Analysis processing, calculates height of water level, significantly reduces manual work cost, also improve the accuracy of measurement.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partially become from the following description
Obviously, or practice through the invention is recognized.
Detailed description of the invention
Above-mentioned and/or additional aspect of the invention and advantage will become from the description of the embodiment in conjunction with the following figures
Obviously and it is readily appreciated that, in which:
Fig. 1 is the flow chart according to the Intelligent water level monitoring based on image recognition of the embodiment of the present invention；
Fig. 2 is the Faster R-CNN network architecture figure according to the embodiment of the present invention；
Fig. 3 is the Faster R-CNN feature extraction network structure according to the embodiment of the present invention；
Fig. 4 is that network and classification Recurrent networks structure are generated according to the candidate region Faster R-CNN of the embodiment of the present invention
Figure；
Fig. 5 is the Grabcut algorithm flow chart according to the embodiment of the present invention；
Fig. 6 is the hardware environment according to the water level monitoring system deployment based on intelligent image identification of the embodiment of the present invention
Figure.
Specific embodiment
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to is used to explain the present invention, and is not considered as limiting the invention.
As shown in Figure 1, the Intelligent water level monitoring method based on image recognition of the embodiment of the present invention, includes the following steps:
Step S1 receives the image data sent by picture pick-up device in tested region, wherein includes water in image data
Ruler image.
Specifically, being equipped with more picture pick-up devices around tested region, the image containing water gauge is shot by picture pick-up device,
Then remote transmission is analyzed and processed to host computer.
Step S2, host computer carry out detection identification to the water gauge in image data using deep neural network, basic herein
On the water gauge image that algorithm is partitioned into the water surface or more cut using figure.
In this step, referring to figs. 2 to Fig. 4, the water gauge in image data is examined using Faster R-CNN algorithm
Survey identification.Faster R-CNN is a classical depth network structure for carrying out target detection, it will be traditional
SelectiveSearch extract mesh calibration method be substituted for network training to realize so that it is full-range detection, classification speed and
Detection accuracy is substantially improved.
Faster R-CNN algorithm is broadly divided into two steps:
(1) it determines the position of target, then the classification of target is identified.In concrete implementation, one is inputted first
Image is opened, is then operated by a series of convolution, pondization, extracts the feature of image, generate characteristic pattern.
(2) network is generated using candidate region, is handled on characteristic pattern, generate the target of different scale and the ratio of width to height
Candidate region.
(3) by characteristic pattern and the object candidate area of generation input classification Recurrent networks, according in object candidate area
Feature exports the classification and object boundary frame of the object candidate area of generation.
Mainly be made of three sub- networks in Faster R-CNN network structure: feature extraction network, candidate region generate
Network and classification Recurrent networks.
Specifically, corresponding to the water gauge image procossing in of the invention, include the following steps:
Firstly, extracting water gauge from the image data received using the feature extraction network of Faster R-CNN algorithm
The feature of image generates characteristic pattern.
Then, it using RPN network as candidate region network, is handled on characteristic pattern, output has a variety of scales
With the rectangular target candidate region of the ratio of width to height.RPN network is that the input of a full convolutional network RPN network is feature extraction network
The characteristic pattern of output exports the rectangular target candidate region with a variety of scales and the ratio of width to height.RPN network passes through one 3 first
× 3 convolution kernel carries out convolution operation to characteristic pattern, forms a feature vector, the convolution kernel for being then 1 × 1 with two sizes
Two full articulamentums are simulated, the score and corrected parameter of candidate region are obtained, normalizing is carried out to score finally by Softmax layers
Change, obtain candidate region whether include object to be measured confidence level.
Finally, needing to carry out candidate region classification after extracting candidate region and returning operation.Classification Recurrent networks
Input is the characteristic pattern of feature extraction network output and the candidate region that network exports is extracted in candidate region, and output is candidate regions
Domain corresponds to the confidence level of each classification and the corrected parameter of candidate region.By characteristic pattern and the input point of the object candidate area of generation
Class Recurrent networks export the classification and water gauge bounding box of the water gauge candidate region of generation according to the feature in candidate region.
In one embodiment of the invention, feature extraction network is a convolutional Neural net in Faster R-CNN algorithm
Network, the network structure can be used as the feature extraction network of Faster R-CNN algorithm, and the characteristics of image of extraction is subsequent network
Input is provided.The feature extraction network of Faster R-CNN algorithm uses following one kind: ZF network, VGG16 network and AlexNet
Network.It should be noted that feature extraction network is not limited to the example above, other kinds of network can also be used, herein not
It repeats again.
It is emerging using feeling according to preset area-of-interest after detecting water gauge image using Faster R-CNN algorithm
The top left co-ordinate and bottom right angular coordinate in interesting region cut the input coordinate of algorithm as figure, cut algorithm from water gauge bounding box using figure
In be partitioned into the water gauge image of the water surface or more, wherein area-of-interest is the default region for confining water gauge.A left side for area-of-interest
It is read in upper angular coordinate slave site water gauge data table, bottom right angular coordinate is consistent with the lower-left angular coordinate of water gauge bounding box.
As shown in figure 5, the figure technology of cutting is the image segmentation algorithm based on graph theory, regard an image as a figure, image
In each pixel represent a node on figure, the relationship between node is regarded side as, is indicated with the similarity between node
The weighted value on its side.During each segmentation, the lesser connection of weight is deleted, so that the higher pixel of similarity is located at
In the same figure, the lower pixel of similarity is located in different figures, is achieved in the continuous division of figure, final to realize to whole
Open the segmentation of image.
The coordinate scaling method of preset area-of-interest is described in detail below:
Due to camera shaking etc., small movement often occurs for the position of water gauge in the picture.Using the figure side of cutting
When method calculates water gauge height, need using area-of-interest top left co-ordinate and bottom right angular coordinate.Bottom right angular coordinate can pass through
Object detection method obtains, and is not required to initialize, top left co-ordinate is initialized." the area-of-interest upper left corner " position one
As take upper left quarter in water gauge, it is ensured that when camera shakes, the X-coordinate and Y-coordinate of water gauge upper left position, which are respectively less than, is equal to sense
The X-coordinate and Y-coordinate of interest region upper left position.
Algorithm Accurate Segmentation is cut using figure and goes out the above water gauge part of the water surface.Faster R-CNN is detected shown in water gauge presence
Water gauge target confines not accurate enough problem, that is, the bounding box and the above water gauge of the water surface of the above water gauge of the water surface detected are in image
In actual boundary frame it is variant.Therefore, the present invention is according further to the area-of-interest (region for confining water gauge) in image,
Algorithm Accurate Segmentation is cut using figure and goes out the above water gauge part of the water surface.Wherein, the top left co-ordinate slave site water gauge of area-of-interest
It is read in tables of data, bottom right angular coordinate and the lower left corner of the target frame detected when application Faster R-CNN detection water gauge are sat
Mark is consistent.Example cuts the segmentation that algorithm carries out the above water gauge part of the water surface to water gauge frame application drawing, and foreground part is the water surface or more
Water gauge part.
Step S3 calculates the water gauge height of the water surface or more according to the water gauge image being partitioned into, and combines the entirety of water gauge
Height, obtain water surface water gauge height below.
Specifically, calculating the water gauge height of the water surface or more according to the water gauge image being partitioned into, include the following steps: basis
Be partitioned into the water gauge image of the water surface or more, using the above water gauge of the water surface pixel number divided by pixel number in unit length (water surface with
Pixel number in pixel number/unit length of upper water gauge), the above water gauge height of the water surface is obtained, then subtracted by the whole height of water gauge
The above water gauge height of the water surface is the following water gauge height of the water surface.Wherein, in unit length pixel number by preset website water gauge number
It is calculated according to table (table 1).
1 website water gauge data table of table
Specifically, the calculation method of " pixel number in unit length " is as follows: in website water gauge data table " in unit length
Pixel number " is needed by being calculated.Calculating process is as follows: the water gauge of certain length is chosen on the image, if its corresponding reality
Border water gauge length is x (can calculate from water gauge scale difference).If the pixel number in the corresponding image of the length is y, unit is long
Pixel number is equal to y/x in spending.
Step S4, the height for obtaining the following water gauge of the water surface obtained in the elevation and step S3 of real site water gauge bottom can
To calculate water level value.
Every technical parameter of the Intelligent water level monitoring method based on image recognition of the embodiment of the present invention is as follows:
(1) use environment
Hardware environment: Windows 10 (Ubuntu 16.4.3), 8G memory (16G memory is excellent), CPU (Intel i7) or
GPU
Software environment: TensorFlow, keras frame, the library OpenCV, python language, MySQL database
(2) software function
1) initialization data library table, database table structure are as shown in table 2；
2 database table structure of table
NODE: the number in identification test river
Position: the water level value of water level when last time tests the website
PixelSise: the pixel number of unit length, unit are how many pixels per cm
Tm: the time when website water level value is tested
X: figure cuts the X-coordinate in the upper left corner for the area-of-interest being applied to when algorithm
Y: figure cuts the Y coordinate data library connection in the upper left corner for the area-of-interest being applied to when algorithm:
Host=" 127.0.0.1 ", user=" root ", passwd=" 123456 ", db=" waterline "
2) request image recognition request (including interface parameters in request, referring in particular to interface specification) is received, to legal figure
As carrying out Intelligent water level detection (including water gauge detection, water level value calculate) or reporting an error to illegal image, detection knot is returned
Fruit etc..
(3) data demand is provided
Data: the image data of offer includes different quality, different application scene, different illumination conditions and different weather feelings
The image of condition.
(4) interface
Network address:
119.3.204.121:8888/delect? stcd=1&imageUrl=http: //www.slhzt.com/img/
1/2.jpg&mId=1&tm=1233
Server address: 119.3.204.121
Port numbers: 8888
Main program entrance: delect
Website number: stcd
Picture address: imageUrl=http: //www.slhzt.com/img/1/2.jpg
No. id: mId of picture
Time: tm
Data transmit http://ip:port/XXXX/XX
It is passed and is joined using post, Transfer Parameters are as shown in table 3.
3 Transfer Parameters table of table
| Attribute type | Attribute-name | Attribute value and its meaning |
| String | token | Token certification |
| String | stcd | Survey station coding |
| String | imageUrl | Picture address |
| String | mId | Picture id |
| Date | tm | Time |
| String | req1 | Required parameter 1 (is stayed and is supplemented) |
| String | req2 | Required parameter 2 (is stayed and is supplemented) |
| String | req3 | Required parameter 3 (is stayed and is supplemented) |
| String | req4 | Required parameter 4 (is stayed and is supplemented) |
Return parameters are shown in Table 4 with the passback of json format, return parameters table.
4 return parameters table of table
(5) it disposes
Water level monitoring system based on intelligent image identification has been deployed on Huawei's Cloud Server of company side's rent, cloud clothes
Device configuration be engaged in as shown in Figure 6.
Intelligent water level monitoring method according to an embodiment of the present invention based on image recognition, the water based on intelligent image identification
Level monitoring system, using relevant computer vision technique and machine learning (including deep learning) algorithm, exploitation based on figure
As the intelligent water gauge reading technology of (video), it can receive what the picture pick-up device teletransmission that reservoir, tunnel etc. are installed came
Image or video containing water gauge, are combined by image processing techniques, deep learning and conventional method, to the water gauge in image
Detected and identified, demarcated the above water gauge part of the water surface, calculate the above water gauge Partial Height of the water surface etc., and then obtain water level number
According to.Realization is automatically saved, shown or is returned request end to calculated waterlevel data.If water level value is more than warning value, can
To carry out water level early warning.The present invention is not necessarily to artificial on-site land survey water level, realizes that image obtains by remote mode, in local progress
Analysis processing, calculates height of water level, significantly reduces manual work cost, also improve the accuracy of measurement.
The performance of the Intelligent water level monitoring method based on image recognition of the embodiment of the present invention describes:
1) situations such as being adapted to different water quality, application scenarios, illumination, weather.It specifically includes:
(1) it is adapted to different water quality situations, such as turbidity, contains floating material etc.；
(2) it is adapted to different application scenarios simultaneously；
(3) illumination condition is needed all to can recognize day and night based on night vision function；
(4) the complicated weather condition of adaptation, such as heavy rainfall, mist, haze are cloudy etc..
2) detection accuracy error+5cm~-5cm.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
Centainly refer to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be any
One or more embodiment or examples in can be combined in any suitable manner.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art are not departing from the principle of the present invention and objective
In the case where can make changes, modifications, alterations, and variations to the above described embodiments within the scope of the invention.The scope of the present invention
By appended claims and its equivalent limit.
Claims (8)
1. a kind of Intelligent water level monitoring method based on image recognition, which comprises the steps of:
Step S1 receives the image data sent by picture pick-up device in tested region, wherein includes water in described image data
Ruler image；
Step S2 carries out detection identification to the water gauge in described image data using deep neural network, uses on this basis
Figure cuts the water gauge image that algorithm is partitioned into the water surface or more；
Step S3 calculates the water gauge height of the water surface or more according to the water gauge image being partitioned into, and combines the whole height of water gauge,
Obtain water surface water gauge height below；
Step S4, the height for obtaining the following water gauge of the water surface obtained in the elevation and step S3 of real site water gauge bottom can be counted
Calculate water level value.
2. the Intelligent water level monitoring method based on image recognition as described in claim 1, which is characterized in that in the step S2
In, detection identification is carried out to the water gauge in described image data using Faster R-CNN algorithm.
3. the Intelligent water level monitoring method based on image recognition as claimed in claim 2, which is characterized in that
The feature extraction network of Faster R-CNN algorithm extracts the feature of water gauge image from the image data received, raw
At characteristic pattern；
It using RPN network as candidate region network, is handled on the characteristic pattern, output has a variety of scales and width high
The rectangular target candidate region of ratio；
By characteristic pattern and the object candidate area of generation input classification Recurrent networks, according to the feature in candidate region, output life
At water gauge candidate region classification and water gauge bounding box.
4. the Intelligent water level monitoring method based on image recognition as claimed in claim 3, which is characterized in that according to preset sense
The input coordinate of algorithm is cut in interest region using the top left co-ordinate of the area-of-interest and with bottom right angular coordinate as figure,
The water gauge image that algorithm is partitioned into the water surface or more is cut using figure.Wherein, the area-of-interest is the default region for confining water gauge.
5. the Intelligent water level monitoring method based on image recognition as claimed in claim 4, which is characterized in that the region of interest
It is read in the top left co-ordinate slave site water gauge data table in domain, the lower-left angular coordinate one of bottom right angular coordinate and the water gauge bounding box
It causes.
6. the Intelligent water level monitoring method based on image recognition as claimed in claim 3, which is characterized in that the Faster
The feature extraction network of R-CNN algorithm uses following one kind: ZF network, VGG16 network and AlexNet network.
7. the Intelligent water level monitoring method based on image recognition as described in claim 1, which is characterized in that in the step S3
In, the water gauge image that the basis is partitioned into calculates the water gauge height of the water surface or more, includes the following steps: to be discharged according to segmentation
Water gauge image more than face obtains the above water of the water surface using the pixel number of the above water gauge of the water surface divided by pixel number in unit length
Ruler height, then subtracting the above water gauge height of the water surface by the whole height of water gauge is the following water gauge height of the water surface.
8. the Intelligent water level monitoring method based on image recognition as claimed in claim 7, which is characterized in that the unit length
Interior pixel number is calculated by preset website water gauge data table.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910517818.3A CN110223341B (en) | 2019-06-14 | 2019-06-14 | Intelligent water level monitoring method based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910517818.3A CN110223341B (en) | 2019-06-14 | 2019-06-14 | Intelligent water level monitoring method based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110223341A true CN110223341A (en) | 2019-09-10 |
| CN110223341B CN110223341B (en) | 2024-05-28 |
Family
ID=67817208
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910517818.3A Active CN110223341B (en) | 2019-06-14 | 2019-06-14 | Intelligent water level monitoring method based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110223341B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110763297A (en) * | 2019-10-09 | 2020-02-07 | 中水淮河规划设计研究有限公司 | Intelligent water gauge water level identification system based on video |
| CN111680606A (en) * | 2020-06-03 | 2020-09-18 | 淮河水利委员会水文局(信息中心) | Low-power-consumption water level remote measuring system based on artificial intelligence cloud identification water gauge |
| CN111967419A (en) * | 2020-08-25 | 2020-11-20 | 深圳大学 | Dam bank dangerous case prediction method and device, computer equipment and storage medium |
| CN112101282A (en) * | 2020-09-25 | 2020-12-18 | 北京瞰天科技有限公司 | Aquatic target identification method and device, electronic equipment and storage medium |
| CN112163496A (en) * | 2020-09-22 | 2021-01-01 | 广东华南水电高新技术开发有限公司 | Embedded terminal reservoir water level early warning method based on semantic segmentation |
| CN112508986A (en) * | 2020-12-04 | 2021-03-16 | 武汉大学 | Water level measurement method based on deep convolutional network and random field |
| CN112523804A (en) * | 2020-12-10 | 2021-03-19 | 广东电网有限责任公司佛山供电局 | Cable tunnel drainage monitoring system and method |
| CN112907506A (en) * | 2021-01-11 | 2021-06-04 | 昆明理工大学 | Water gauge color information-based indefinite-length water gauge water level detection method and device and storage medium |
| CN113223073A (en) * | 2021-04-16 | 2021-08-06 | 北京科技大学 | River channel water level assessment method |
| CN113819971A (en) * | 2020-07-07 | 2021-12-21 | 湖北亿立能科技股份有限公司 | Artificial intelligence water level monitoring system based on water, scale and floater semantic segmentation |
| CN114593792A (en) * | 2022-03-29 | 2022-06-07 | 中国水利水电科学研究院 | Underground water level monitoring method and device and storage medium |
| CN115457563A (en) * | 2022-08-24 | 2022-12-09 | 浙江工业大学 | Zero-missing-detection and low-error-identification ship water gauge reading method |
| CN118566223A (en) * | 2024-06-24 | 2024-08-30 | 中控技术股份有限公司 | A method for measuring alcohol content online |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105760886A (en) * | 2016-02-23 | 2016-07-13 | 北京联合大学 | Image scene multi-object segmentation method based on target identification and saliency detection |
| CN106504233A (en) * | 2016-10-18 | 2017-03-15 | 国网山东省电力公司电力科学研究院 | Image electric power widget recognition methodss and system are patrolled and examined based on the unmanned plane of Faster R CNN |
| CN107367310A (en) * | 2017-07-11 | 2017-11-21 | 华南理工大学 | A kind of river level remote monitoring method based on computer vision |
| CN107563303A (en) * | 2017-08-09 | 2018-01-09 | 中国科学院大学 | A kind of robustness Ship Target Detection method based on deep learning |
| US20180122082A1 (en) * | 2016-11-02 | 2018-05-03 | General Electric Company | Automated segmentation using deep learned priors |
| CN108073905A (en) * | 2017-12-21 | 2018-05-25 | 北京邮电大学 | A kind of method, system and the equipment of intelligence water gauge reading |
| CN108509919A (en) * | 2018-04-03 | 2018-09-07 | 哈尔滨哈船智控科技有限责任公司 | A kind of detection and recognition methods based on deep learning to waterline in video or picture |
| CN108596221A (en) * | 2018-04-10 | 2018-09-28 | 江河瑞通（北京）技术有限公司 | The image-recognizing method and equipment of rod reading |
| CN108648233A (en) * | 2018-03-24 | 2018-10-12 | 北京工业大学 | A kind of target identification based on deep learning and crawl localization method |
| CN108985322A (en) * | 2018-06-01 | 2018-12-11 | 广东电网有限责任公司 | A kind of cable tunnel ponding positioning identifying method based on ZF-Faster RCNNs |
| CN109063586A (en) * | 2018-07-11 | 2018-12-21 | 东南大学 | A kind of Faster R-CNN driver's detection method based on candidate's optimization |
| CN109145830A (en) * | 2018-08-24 | 2019-01-04 | 浙江大学 | A kind of intelligence water gauge recognition methods |
| KR101970442B1 (en) * | 2018-12-04 | 2019-04-19 | 주식회사 넥스파시스템 | Illegal parking enforcement system Using Fast R-CNN based on Vehicle detection |
| KR20190063729A (en) * | 2017-11-30 | 2019-06-10 | 주식회사 넥토마이닝 | Life protection system for social disaster using convergence technology like camera, sensor network, and directional speaker system |
-
2019
- 2019-06-14 CN CN201910517818.3A patent/CN110223341B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105760886A (en) * | 2016-02-23 | 2016-07-13 | 北京联合大学 | Image scene multi-object segmentation method based on target identification and saliency detection |
| CN106504233A (en) * | 2016-10-18 | 2017-03-15 | 国网山东省电力公司电力科学研究院 | Image electric power widget recognition methodss and system are patrolled and examined based on the unmanned plane of Faster R CNN |
| US20180122082A1 (en) * | 2016-11-02 | 2018-05-03 | General Electric Company | Automated segmentation using deep learned priors |
| CN107367310A (en) * | 2017-07-11 | 2017-11-21 | 华南理工大学 | A kind of river level remote monitoring method based on computer vision |
| CN107563303A (en) * | 2017-08-09 | 2018-01-09 | 中国科学院大学 | A kind of robustness Ship Target Detection method based on deep learning |
| KR20190063729A (en) * | 2017-11-30 | 2019-06-10 | 주식회사 넥토마이닝 | Life protection system for social disaster using convergence technology like camera, sensor network, and directional speaker system |
| CN108073905A (en) * | 2017-12-21 | 2018-05-25 | 北京邮电大学 | A kind of method, system and the equipment of intelligence water gauge reading |
| CN108648233A (en) * | 2018-03-24 | 2018-10-12 | 北京工业大学 | A kind of target identification based on deep learning and crawl localization method |
| CN108509919A (en) * | 2018-04-03 | 2018-09-07 | 哈尔滨哈船智控科技有限责任公司 | A kind of detection and recognition methods based on deep learning to waterline in video or picture |
| CN108596221A (en) * | 2018-04-10 | 2018-09-28 | 江河瑞通（北京）技术有限公司 | The image-recognizing method and equipment of rod reading |
| CN108985322A (en) * | 2018-06-01 | 2018-12-11 | 广东电网有限责任公司 | A kind of cable tunnel ponding positioning identifying method based on ZF-Faster RCNNs |
| CN109063586A (en) * | 2018-07-11 | 2018-12-21 | 东南大学 | A kind of Faster R-CNN driver's detection method based on candidate's optimization |
| CN109145830A (en) * | 2018-08-24 | 2019-01-04 | 浙江大学 | A kind of intelligence water gauge recognition methods |
| KR101970442B1 (en) * | 2018-12-04 | 2019-04-19 | 주식회사 넥스파시스템 | Illegal parking enforcement system Using Fast R-CNN based on Vehicle detection |
Non-Patent Citations (2)
| Title |
|---|
| 郭文生;包灵;钱智成;曹万里;: "基于自适应叠合分割与深度神经网络的人数统计方法", 计算机科学, no. 08 * |
| 黄劲潮;: "基于快速区域建议网络的图像多目标分割算法", 山东大学学报(工学版), no. 04, 25 May 2018 (2018-05-25) * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110763297A (en) * | 2019-10-09 | 2020-02-07 | 中水淮河规划设计研究有限公司 | Intelligent water gauge water level identification system based on video |
| CN111680606A (en) * | 2020-06-03 | 2020-09-18 | 淮河水利委员会水文局(信息中心) | Low-power-consumption water level remote measuring system based on artificial intelligence cloud identification water gauge |
| CN113819971A (en) * | 2020-07-07 | 2021-12-21 | 湖北亿立能科技股份有限公司 | Artificial intelligence water level monitoring system based on water, scale and floater semantic segmentation |
| CN111967419B (en) * | 2020-08-25 | 2023-06-20 | 深圳大学 | Dam bank dangerous case prediction method, dam bank dangerous case prediction device, computer equipment and storage medium |
| CN111967419A (en) * | 2020-08-25 | 2020-11-20 | 深圳大学 | Dam bank dangerous case prediction method and device, computer equipment and storage medium |
| CN112163496A (en) * | 2020-09-22 | 2021-01-01 | 广东华南水电高新技术开发有限公司 | Embedded terminal reservoir water level early warning method based on semantic segmentation |
| CN112163496B (en) * | 2020-09-22 | 2024-05-28 | 广东华南水电高新技术开发有限公司 | Embedded terminal reservoir water level early warning method based on semantic segmentation |
| CN112101282A (en) * | 2020-09-25 | 2020-12-18 | 北京瞰天科技有限公司 | Aquatic target identification method and device, electronic equipment and storage medium |
| CN112101282B (en) * | 2020-09-25 | 2024-04-26 | 北京瞰天科技有限公司 | Water target identification method and device, electronic equipment and storage medium |
| CN112508986A (en) * | 2020-12-04 | 2021-03-16 | 武汉大学 | Water level measurement method based on deep convolutional network and random field |
| CN112523804A (en) * | 2020-12-10 | 2021-03-19 | 广东电网有限责任公司佛山供电局 | Cable tunnel drainage monitoring system and method |
| CN112907506A (en) * | 2021-01-11 | 2021-06-04 | 昆明理工大学 | Water gauge color information-based indefinite-length water gauge water level detection method and device and storage medium |
| CN113223073A (en) * | 2021-04-16 | 2021-08-06 | 北京科技大学 | River channel water level assessment method |
| CN113223073B (en) * | 2021-04-16 | 2022-04-19 | 北京科技大学 | River channel water level assessment method |
| CN114593792A (en) * | 2022-03-29 | 2022-06-07 | 中国水利水电科学研究院 | Underground water level monitoring method and device and storage medium |
| CN115457563A (en) * | 2022-08-24 | 2022-12-09 | 浙江工业大学 | Zero-missing-detection and low-error-identification ship water gauge reading method |
| CN118566223A (en) * | 2024-06-24 | 2024-08-30 | 中控技术股份有限公司 | A method for measuring alcohol content online |
| CN118566223B (en) * | 2024-06-24 | 2025-01-14 | 中控技术股份有限公司 | A method for measuring alcohol content online |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110223341B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110223341A (en) | A kind of Intelligent water level monitoring method based on image recognition | |
| Cenci et al. | Integrating remote sensing and GIS techniques for monitoring and modeling shoreline evolution to support coastal risk management | |
| Topouzelis et al. | Detection and discrimination between oil spills and look-alike phenomena through neural networks | |
| Kalke et al. | Support vector machine learning applied to digital images of river ice conditions | |
| Benassai et al. | Rip current evidence by hydrodynamic simulations, bathymetric surveys and UAV observation | |
| Kumar et al. | Mapping shoreline change using machine learning: a case study from the eastern Indian coast | |
| Mucerino et al. | Coastal exposure assessment on Bonassola bay | |
| de Santiago et al. | A comparative study of models to predict storm impact on beaches | |
| Benassai et al. | Sea wave modeling with X-band COSMO-SkyMed© SAR-derived wind field forcing and applications in coastal vulnerability assessment | |
| Ren et al. | Detection of SST fronts from a high-resolution model and its preliminary results in the south China sea | |
| Yoshida et al. | Estimation of distributed flow resistance in vegetated rivers using airborne topo‐bathymetric LiDAR and its application to risk management tasks for Asahi River flooding | |
| Liu et al. | Lifeguarding Operational Camera Kiosk System (LOCKS) for flash rip warning: Development and application | |
| WO2024153703A1 (en) | Flood segmentation method using synthetic aperture radar images and digital elevation models | |
| CN113870224A (en) | Flood monitoring method, system, equipment and medium | |
| Tillihal et al. | River flood monitoring and management: a review | |
| Bhunia | Assessment of automatic extraction of surface water dynamism using multi-temporal satellite data | |
| CN112380984A (en) | A spatial evaluation method for slow-flow capacity of salt marsh vegetation based on remote sensing | |
| Goncalves et al. | Identification, characterization and analysis of the Douro river plume from MERIS data | |
| Zeng et al. | Urban water body detection from the combination of high-resolution optical and SAR images | |
| Chang et al. | Multi-lens pier scour monitoring and scour depth prediction | |
| Emami et al. | Spatial patterns in groundwater seepage and surf zone morphology: Muriwai Beach, New Zealand | |
| Aquilino et al. | Multi-temporal land use analysis of an ephemeral river area using an artificial neural network approach on Landsat imagery | |
| Vishwakarma et al. | Assessment of reservoir sedimentation using remote sensing technique with GIS model-A review | |
| Farnoush et al. | Discharge estimation and rating curve derivation, using satellite geometry data and isovel contours at Karun River, Iran | |
| Casado et al. | Emerging remote sensing technologies for flood applications |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |