CN111476120B - Unmanned aerial vehicle intelligent ship water gauge identification method and device - Google Patents
Unmanned aerial vehicle intelligent ship water gauge identification method and device Download PDFInfo
- Publication number
- CN111476120B CN111476120B CN202010222648.9A CN202010222648A CN111476120B CN 111476120 B CN111476120 B CN 111476120B CN 202010222648 A CN202010222648 A CN 202010222648A CN 111476120 B CN111476120 B CN 111476120B
- Authority
- CN
- China
- Prior art keywords
- water gauge
- image
- aerial vehicle
- unmanned aerial
- waterline
- 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.)
- Active
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 164
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000013528 artificial neural network Methods 0.000 claims abstract description 21
- 230000011218 segmentation Effects 0.000 claims abstract description 6
- 238000012545 processing Methods 0.000 claims description 19
- 230000008859 change Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 11
- 230000009467 reduction Effects 0.000 claims description 10
- 239000000049 pigment Substances 0.000 claims description 7
- 238000009499 grossing Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 230000007797 corrosion Effects 0.000 claims description 2
- 238000005260 corrosion Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 5
- 230000008569 process Effects 0.000 abstract description 5
- 238000005259 measurement Methods 0.000 abstract description 3
- 238000001514 detection method Methods 0.000 description 8
- 230000008901 benefit Effects 0.000 description 4
- 230000008676 import Effects 0.000 description 4
- 239000003245 coal Substances 0.000 description 3
- 230000000630 rising effect Effects 0.000 description 3
- 239000002699 waste material Substances 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000005303 weighing Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000002238 attenuated effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000000969 carrier Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000010339 dilation Effects 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 239000013535 sea water Substances 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
- G06V20/13—Satellite images
-
- 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
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Astronomy & Astrophysics (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Remote Sensing (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Measurement Of Levels Of Liquids Or Fluent Solid Materials (AREA)
Abstract
The invention relates to the technical field of water gauge identification, and discloses a method and a device for identifying an intelligent ship water gauge of an unmanned aerial vehicle and a computer storage medium, wherein the method comprises the following steps: acquiring an original image of a ship water gauge through an unmanned aerial vehicle; separating the ship body and the water surface in the original image by using the color difference between the ship body and the water surface and adopting a color gamut segmentation method to obtain the position of a waterline; intercepting the area where the water gauge numbers are located in the original image as a candidate area, and identifying all the water gauge numbers in the candidate area through a digital identification model obtained through neural network training; and carrying out water gauge identification according to the relative position relation between the waterline position and the water gauge number. The method for identifying the water gauge of the intelligent ship of the unmanned aerial vehicle has the technical effects of safe measurement process and high measurement precision.
Description
Technical Field
The invention relates to the technical field of water gauge identification, in particular to a method and a device for identifying an intelligent ship water gauge of an unmanned aerial vehicle and a computer storage medium.
Background
With the development of global economic trade, as an important method widely used for ship transportation goods metering and import and export commodity weight identification at home and abroad, the water gauge weighing based on the Archimedes principle shows more and more important status, and the reading accuracy of ship draft numerical values becomes the most key factor of the water gauge weighing accuracy.
Data show that the finished cargo throughput of the ports in China in 2017 is 1.2644 multiplied by 1011t, the concordance increase is 6.4 percent, wherein the concordance acceleration of the coastal port and the inland river port is basically equal to 6.4 percent and 6.3 percent respectively, and the finished cargo throughput is 8.625 multiplied by 10 respectively9t and 4.019X 109t; 4.002 x 10 of foreign trade cargo throughput is completed9t, increase by 5.7% on the same scale. The import quantity of bulk foreign trade is rapidly increased, the throughput speed of coal is accelerated, wherein the import quantity of the foreign trade coal is increased unexpectedly, the import quantity is greatly increased, the main trade transportation mode of the bulk goods is ship water transportation, and along with the development of the ship transportation industry, the ship cargo weight identification is taken as the handing-over junction of the goods in the bulk tradeThe accuracy of ship cargo weight appraisal is receiving more and more attention from people according to the basis of calculating and processing claims, calculating freight, closing and tax, and the like. The ship cargo weight identification can be used as a proof of delivery or settlement when the ship cargo weight identification is exported, and can be used as a basis for landed pricing or short and heavy claims when the ship cargo weight identification is imported, and an accurate metering result has extremely important significance for protecting the benefits of carriers, shippers and receivers. Therefore, in the trade process of bulk goods such as coal in China, accurate identification of trade volume is more important, and accurate judgment of the water gauge draught value of a ship is the basis.
The existing water gauge value detection methods are various, and most representative methods are manual observation, pressure sensor detection and sonar detection.
(1) And (3) manual observation: because the manual observation is disturbed by many aspects of subjective and objective factors of people, even many people observe the water gauge, final result is also not too scientific, and when the surface of water fluctuation is bigger, not only can the detection error still increase once more, also probably leads to measurement personnel's life to suffer the threat simultaneously, except this, general detection needs to lease the boat, is close to boats and ships and detects, not only is not cost-effective in economy, has great manpower and time waste moreover.
(2) Detecting by a pressure sensor: the pressure sensor mainly utilizes the pressure sensor that the steamer hull outside was adorned in advance, goes to obtain the depth of water change through the pressure to carrying cargo the front and back, and the precision is higher, but because instrument precision is high, the installation is inconvenient, later stage need be maintained, and is related to sea water density, difficult popularization, practical value is not big.
(3) Sonar detection: sonar detection is achieved by using the characteristic that the sound energy of ultrasonic waves is attenuated little in water, but has the same disadvantage as pressure sensor detection.
Disclosure of Invention
The invention aims to overcome the technical defects, provides a method and a device for identifying a water gauge of an intelligent ship of an unmanned aerial vehicle and a computer storage medium, and solves the technical problems of low water gauge identification precision and potential safety hazard in the prior art.
In order to achieve the technical purpose, the technical scheme of the invention provides an unmanned aerial vehicle intelligent ship water gauge identification method, which comprises the following steps:
acquiring an original image of a ship water gauge through an unmanned aerial vehicle;
separating the ship body and the water surface in the original image by using the color difference between the ship body and the water surface and adopting a color gamut segmentation method to obtain the position of a waterline;
intercepting the area where the water gauge numbers are located in the original image as a candidate area, and identifying all the water gauge numbers in the candidate area through a digital identification model obtained through neural network training;
and carrying out water gauge identification according to the relative position relation between the waterline position and the water gauge number.
The invention also provides an unmanned aerial vehicle intelligent ship water gauge identification device which comprises a processor and a memory, wherein the memory is stored with a computer program, and the computer program is executed by the processor to realize the unmanned aerial vehicle intelligent ship water gauge identification method.
The invention also provides a computer storage medium, wherein a computer program is stored on the computer storage medium, and when the computer program is executed by a processor, the unmanned aerial vehicle intelligent ship water gauge identification method is realized.
Compared with the prior art, the invention has the beneficial effects that: according to the invention, the technology of a neural network in the aspect of computer vision is combined with the unmanned aerial vehicle technology, firstly, relevant water gauge image information of a ship is rapidly acquired by the unmanned aerial vehicle at a port and a river basin with complex environment, and the acquired water gauge image information is returned to a server for intelligent identification, so that the characteristics of high maneuverability and high zooming of the unmanned aerial vehicle are utilized, the waste of financial resources is reduced, the human resources are released, and the life safety of customs personnel is guaranteed; meanwhile, after the original image is subjected to waterline recognition, the water gauge number is automatically recognized by utilizing the neural network, the recognition precision is high, manual recognition is not needed, and the occurrence of greedy pollution in the aspect of the water gauge is fundamentally avoided.
Drawings
Fig. 1 is a flowchart of an embodiment of a method for identifying a water gauge of an intelligent ship of an unmanned aerial vehicle, provided by the invention;
fig. 2 is a schematic view of water line identification in an embodiment of the unmanned aerial vehicle intelligent ship water gauge identification method provided by the invention;
fig. 3 is a neural network structure diagram of an embodiment of the unmanned aerial vehicle intelligent ship water gauge identification method provided by the invention;
fig. 4 is a schematic view of water level reading of an embodiment of the unmanned aerial vehicle intelligent ship water gauge identification method provided by the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
As shown in fig. 1, embodiment 1 of the present invention provides a method for identifying a water gauge of an intelligent ship of an unmanned aerial vehicle, which is hereinafter referred to as the method, and includes the following steps:
s1, acquiring an original image of the ship water gauge through an unmanned aerial vehicle;
s2, separating the ship body and the water surface in the original image by using the color difference between the ship body and the water surface and adopting a color gamut segmentation method to obtain a waterline position;
s3, intercepting the area where the water gauge numbers are located in the original image as a candidate area, and identifying all the water gauge numbers in the candidate area through a digital identification model obtained through neural network training;
and S4, carrying out water gauge identification according to the relative position relation between the waterline position and the water gauge number.
This embodiment utilizes unmanned aerial vehicle to acquire when acquireing the original image of water gauge, and unmanned aerial vehicle has the advantage that can acquire the image fast at the harbour that the environment is complicated and basin, and unmanned aerial vehicle gathers and carries out intelligent recognition with image return server after accomplishing, can return the image here through wireless transmission's mode to improve convenience, the real-time of image acquisition and passback. After the original image is collected, firstly, waterline recognition is carried out on the original image by utilizing the color difference between the ship body and the water surface, after the waterline recognition is carried out on the original image, automatic recognition is carried out on the water gauge number by utilizing the neural network, the recognition precision is high, and manual recognition is not needed. And after the water line and the number are identified, the water gauge can be read. The invention utilizes the characteristics of high maneuverability and high zooming of the unmanned aerial vehicle, realizes the acquisition of the original image of the water gauge safely, efficiently and at low cost, reduces the waste of financial resources, releases human resources and ensures the life safety of customs personnel; meanwhile, automatic identification of numbers on the water gauge is realized through a neural network, and the occurrence of greedy pollution in the aspect of the water gauge is fundamentally avoided.
Preferably, the color difference between the ship body and the water surface is utilized, the ship body and the water surface in the original image are separated by adopting a color gamut segmentation method, and the waterline position is obtained, and specifically:
converting the original image into a gray image, performing noise reduction processing on the gray image, and performing pixel value traversal on the gray image subjected to noise reduction to obtain a first waterline position;
converting the original image into an HSV image, taking the average value of pigments in a set area in the HSV image as an anchor, finding a change area with the maximum pigment change by sliding the anchor in the HSV image, obtaining a vector of the change area by using a minAreaRect method in OpenCV, removing the change area with the vector angle within a certain angle range, and then selecting the change area according with the waterline characteristics according to the length-width ratio to judge and obtain the position of a second waterline;
and fitting according to the first waterline position and the second waterline position to obtain the waterline position.
The core of waterline discernment is that there is the colour difference between utilization hull and the surface of water, goes to separate hull and surface of water through the mode that the colour gamut was cut apart, in order to improve the accuracy of waterline discernment, the discernment of waterline in this embodiment mainly starts with from two kinds of picture processing: firstly, after an original image is converted into a gray image, noise reduction processing is carried out, the gray image subjected to noise reduction is traversed by utilizing an image pixel value, and the position of a waterline in the image is obtained. There are many methods for pixel value traversal, for example, OpenCV has a function of multiple pixel values traversed, and users can select the function according to requirements. Secondly, after the original image is converted into an HSV image, the average value of pigments in a certain area of the HSV image is taken, then the anchor is used for sliding in an image window to find an area with extremely changed pigments, and then the position of a waterline is judged according to the vector and the area of the pigments in the area. Specifically, the certain angle range refers to an angle range in which an included angle between a vector angle and a waterline angle is close to a right angle, the length-width ratio refers to a length-width ratio which meets the characteristics of the waterline, namely, a length-width ratio of a strip-shaped line, and the specific length-width ratio can be set according to actual conditions, for example, 200: 1. And finally, combining the result obtained by the first processing and the result obtained by the second processing to fit a more accurate waterline height. Specifically, as shown in fig. 2, two rows of dots in fig. 2 are waterline positions obtained by processing one and processing two, respectively, and a straight line between the two rows of dots is the waterline position obtained by fitting.
Preferably, the noise reduction processing is performed on the grayscale image, specifically:
carrying out noise reduction smoothing processing on the gray level image by using a Canny operator;
carrying out expansion corrosion treatment on the noise-reduced and smoothed gray level image;
and further denoising the expanded and corroded gray level image by using a findContours method in OpenCV to obtain a denoised gray level image.
In order to avoid the influence of image noise on waterline recognition, the image needs to be denoised first, specifically, after an original image is converted into a single-channel gray image, a Canny operator is used for denoising and smoothing the image, and when a single-channel histogram of the image is researched, the R channel is best in extraction effect when double thresholds of Canny are between 18 and 54, therefore, in the embodiment, the thresholds of the Canny operator are selected from 18-54, then the processed image is subjected to dilation and erosion processing to eliminate fine noise, and finally functions such as findContours in an OpenCV library are used for further eliminating unnecessary noise.
Preferably, the area where the footage number is located in the original image is intercepted as a candidate area, specifically:
carrying out gray processing on the original image to obtain a gray image, and extracting an MSER (minimum shift error rate) region in the gray image by adopting a maximum stable extremum region algorithm;
and denoising and de-duplicating the MSER region to obtain the candidate region.
The number identification of the water gauge is also divided into two parts, the first part being a candidate area for determining the number of the water gauge in the image, and the second part being the number in the identified candidate area. The first part of this embodiment uses the maximum stable extremum region algorithm based on the watershed algorithm, namely MSER. The maximum stable extremum region algorithm is specifically that the gray level image is subjected to binarization processing, a binarization threshold value is [0,255], and the binarization threshold value is sequentially increased from 0 to 255, so that the binarization image is subjected to a process from full black to full white (as an overhead view image with the water level rising continuously), in the process, the areas of some connected regions change rapidly along with the rising of the threshold value, namely, unstable regions are submerged in the unstable regions in the process, the areas of some connected regions change little along with the rising of the threshold value, and the regions are called MSER and the MSER regions are stored. Then, the height and width of the storage area and the area are judged, so that the noise area in the MSER area and the same block area of the repeated mark are eliminated.
Preferably, the method for identifying all water gauge numbers in the candidate region by using a digital identification model obtained by neural network training specifically comprises the following steps:
collecting and preprocessing a water gauge image, establishing a sample data set, and carrying out digital marking on the water gauge image in the sample data set;
training the neural network by taking the marked water gauge image as a training sample to obtain a water gauge digital recognition model;
and identifying the water gauge number in the candidate area through the number identification model.
The sample data sources for neural network training in this embodiment are: a professional uses the unmanned aerial vehicle to shoot videos including a water gauge of a ship to be detected at multiple angles and distances, and key frames in the videos are intercepted and used as training data of a neural network. Data preprocessing: and screening invalid samples with the sample data set including blur, obvious noise and no target to be detected, ensuring that each sample picture clearly contains the characteristic data to be detected, and performing different processing due to different requirements of each link on the characteristic points of the image. After the samples are collected, a neural network is built to identify numbers of candidate areas, a neural network backbone in this embodiment is shown in fig. 3, and includes an input layer (input), a convolution layer (convolution), a pooling layer (maxpool), and a full connection layer (full connected), wherein the full connection layer is added with a droout layer to prevent overfitting, a loss function is a cross entropy loss function, namely, a preferential _ cross entropy, a learning rate is 0.001, an optimizer uses a first-order optimization algorithm Adam, and an activation function (activation) is relu.
Preferably, the method for training the neural network by using the marked water gauge image as a training sample to obtain the digital recognition model further comprises:
training a neural network by using a Google street view house number data set to obtain a pre-training model;
and training the pre-training model by using the marked water gauge image to obtain the digital recognition model.
In order to improve the accuracy of the initial model, the patent firstly trains by using an svhn Google streetscape number data set provided by Google to obtain a pre-training model. In order to match with the pre-training model, the water gauge image needs to be converted into the size same as the size of the picture in the svhn *** streetscape number data set, namely the size of 32 × 32, 3 channels, and the label corresponding to the water gauge image is recorded in a one _ hot coding format.
Preferably, the water gauge is identified according to the relative position relationship between the waterline position and the water gauge number, specifically:
fitting according to the pixel relative position relation of each water gauge number to obtain a curvature function of a fitting curve;
determining the pixel coordinates of the waterline in the original image according to the waterline position and the curvature function;
acquiring the water gauge numbers of different units positioned at the lowest part in all the water gauge numbers to obtain unit water gauge numbers;
and determining water level readings according to the position relation between each unit number and the waterline coordinate.
After the position of the waterline is confirmed and the number of the water gauge is recognized, automatic reading is needed. Specifically, each water scale number is used as a node, the pixel position of each node is subjected to mathematical fitting, the curvature function of a fitting line is obtained, and the corresponding relation between the pixel position and the water scale number is obtained, so that the water scale can be understood as the scale of the image. And determining the pixel coordinates of the waterline in the image according to the waterline position and the curvature. Acquiring the water gauge numbers of different units positioned at the bottom in all the water gauge numbers to obtain the unit water gauge numbers, specifically, as shown in fig. 4, the water gauge units in the embodiment include meters and centimeters, the number of meters in water depth is judged by using the M symbol at the bottom and the number on the left of M in water gauge number identification, namely, the maximum number "8" in fig. 4, and when the position is not clear, the number on the left of M on the upper side is searched; the number of centimeters near the water level is judged by using the lowest number in the water gauge numbers, such as 8.20, 8.10 and 7.90 in fig. 4. And then, determining a water level reading according to the mapping relation between the water gauge number and the pixel and the height, wherein for example, a waterline is between 8.20 and 8.10, the water gauge reading is between 8.10 and 8.20M, and the centimeter number is estimated according to the waterline coordinate to obtain a final reading, for example, 8.18.
Preferably, the method further comprises:
acquiring an original video of a ship water gauge for a period of time through an unmanned aerial vehicle, and intercepting a plurality of frames of original images from the original video;
and respectively carrying out water gauge identification on the basis of each frame of the original image to obtain corresponding water level readings, screening the mode of the water level readings, and solving the average value of the mode to obtain the final water level reading.
Because the surface of water is unstable usually, the hull can influence the residual error of image and the accurate discernment of water gauge along with rocking of surface of water, consequently gathers the original video of water gauge of a period of time through unmanned aerial vehicle, intercepts multiframe original image, carries out above-mentioned processing to each frame original image and obtains its corresponding water level reading, screens the mode of each water level reading, then seeks the average value, obtains more accurate water level reading.
Example 2
The embodiment 2 of the invention provides an unmanned aerial vehicle intelligent ship water gauge recognition device which comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the unmanned aerial vehicle intelligent ship water gauge recognition method provided by the embodiment 1 is realized.
The intelligent ship water gauge identification device for the unmanned aerial vehicle is used for realizing the intelligent ship water gauge identification method for the unmanned aerial vehicle, so that the intelligent ship water gauge identification method for the unmanned aerial vehicle has the technical effects, and the intelligent ship water gauge identification device for the unmanned aerial vehicle also has the advantages, and the details are not repeated herein.
Example 3
The computer storage medium provided by the embodiment of the invention is used for realizing the unmanned aerial vehicle intelligent ship water gauge identification method, so that the unmanned aerial vehicle intelligent ship water gauge identification method has the technical effects, and the computer storage medium also has the technical effects, and is not repeated herein.
The above-described embodiments of the present invention should not be construed as limiting the scope of the present invention. Any other corresponding changes and modifications made according to the technical idea of the present invention should be included in the protection scope of the claims of the present invention.
Claims (9)
1. An unmanned aerial vehicle intelligent ship water gauge identification method is characterized by comprising the following steps:
acquiring an original image of a ship water gauge through an unmanned aerial vehicle;
separating the ship body and the water surface in the original image by using the color difference between the ship body and the water surface and adopting a color gamut segmentation method to obtain the position of a waterline;
intercepting the area where the water gauge numbers are located in the original image as a candidate area, and identifying all the water gauge numbers in the candidate area through a digital identification model obtained through neural network training;
carrying out water gauge identification according to the relative position relation between the waterline position and the water gauge number;
separating the hull and the water surface in the original image by using the color difference between the hull and the water surface and adopting a color gamut segmentation method to obtain the position of a waterline, which specifically comprises the following steps:
converting the original image into a gray image, performing noise reduction processing on the gray image, and performing pixel value traversal on the gray image subjected to noise reduction to obtain a first waterline position;
converting the original image into an HSV image, taking the average value of pigments in a set area in the HSV image as an anchor, searching for a change area with the maximum pigment change by utilizing the anchor to slide in the HSV image, obtaining a vector of the change area, removing the change area with the vector angle within a certain angle range, then selecting the change area which accords with the waterline characteristics according to the length-width ratio, and judging to obtain the position of a second waterline;
and fitting according to the first waterline position and the second waterline position to obtain the waterline position.
2. The unmanned aerial vehicle intelligent ship water gauge identification method according to claim 1, wherein the grayscale image is subjected to noise reduction processing, specifically:
carrying out noise reduction smoothing processing on the gray level image by using a Canny operator;
carrying out expansion corrosion treatment on the noise-reduced and smoothed gray level image;
and further denoising the expanded and corroded gray level image by using a findContours method in OpenCV to obtain a denoised gray level image.
3. The unmanned aerial vehicle intelligent ship water gauge identification method according to claim 1, wherein the area where the water gauge numbers are located in the original image is intercepted as a candidate area, and specifically:
carrying out gray processing on the original image to obtain a gray image, and extracting an MSER (minimum shift error rate) region in the gray image by adopting a maximum stable extremum region algorithm;
and denoising and de-duplicating the MSER region to obtain the candidate region.
4. The unmanned aerial vehicle intelligent ship water gauge identification method according to claim 1, wherein a digital identification model obtained through neural network training identifies all water gauge numbers in the candidate area, specifically:
collecting and preprocessing a water gauge image, establishing a sample data set, and carrying out digital marking on the water gauge image in the sample data set;
training the neural network by taking the marked water gauge image as a training sample to obtain a water gauge digital recognition model;
and identifying the water gauge number in the candidate area through the number identification model.
5. The unmanned aerial vehicle intelligent ship water gauge identification method according to claim 1, wherein the marked water gauge image is used as a training sample to train a neural network to obtain a digital identification model, and the method further comprises the following steps:
training a neural network by using a Google street view house number data set to obtain a pre-training model;
and training the pre-training model by using the marked water gauge image to obtain the digital recognition model.
6. The unmanned aerial vehicle intelligent ship water gauge identification method according to claim 1, wherein the water gauge identification is carried out according to the relative position relation between the waterline position and the water gauge number, and specifically comprises the following steps:
fitting according to the pixel relative position relation of each water gauge number to obtain a curvature function of a fitting curve;
determining the pixel coordinates of the waterline in the original image according to the waterline position and the curvature function;
acquiring the water gauge numbers of different units positioned at the lowest part in all the water gauge numbers to obtain unit water gauge numbers;
and determining water level readings according to the position relation between each unit number and the waterline coordinate.
7. The unmanned aerial vehicle smart ship water gauge identification method according to claim 1, further comprising:
acquiring an original video of a ship water gauge for a period of time through an unmanned aerial vehicle, and intercepting a plurality of frames of original images from the original video;
and respectively carrying out water gauge identification on the basis of each frame of the original image to obtain corresponding water level readings, screening the mode of the water level readings, and solving the average value of the mode to obtain the final water level reading.
8. An unmanned aerial vehicle intelligent ship water gauge identification device, which comprises a processor and a memory, wherein the memory is stored with a computer program, and when the computer program is executed by the processor, the unmanned aerial vehicle intelligent ship water gauge identification device realizes the unmanned aerial vehicle intelligent ship water gauge identification method according to any one of claims 1-7.
9. A computer storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the drone smart vessel water gauge identification method of any of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010222648.9A CN111476120B (en) | 2020-03-26 | 2020-03-26 | Unmanned aerial vehicle intelligent ship water gauge identification method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010222648.9A CN111476120B (en) | 2020-03-26 | 2020-03-26 | Unmanned aerial vehicle intelligent ship water gauge identification method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111476120A CN111476120A (en) | 2020-07-31 |
CN111476120B true CN111476120B (en) | 2021-03-23 |
Family
ID=71750235
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010222648.9A Active CN111476120B (en) | 2020-03-26 | 2020-03-26 | Unmanned aerial vehicle intelligent ship water gauge identification method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111476120B (en) |
Families Citing this family (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112183470B (en) * | 2020-10-28 | 2022-07-19 | 长江大学 | Ship water gauge identification method and equipment and storage medium |
CN112308009B (en) * | 2020-11-12 | 2024-02-27 | 湖北九感科技有限公司 | Water gauge water level identification method and device |
CN112487987A (en) * | 2020-11-30 | 2021-03-12 | 江苏云控软件技术有限公司 | Method for measuring boundary between ship and water surface |
CN112862820B (en) * | 2021-03-17 | 2023-12-29 | 水利部交通运输部国家能源局南京水利科学研究院 | Image recognition technology-based intelligent acquisition device and method for water depth of manometric tube group |
CN115439861A (en) * | 2022-09-30 | 2022-12-06 | 北京中盛益华科技有限公司 | Water gauge recognition method based on OCR |
CN117197048B (en) * | 2023-08-15 | 2024-03-08 | 力鸿检验集团有限公司 | Ship water gauge reading detection method, device and equipment |
CN117251943B (en) * | 2023-11-20 | 2024-02-06 | 力鸿检验集团有限公司 | Waterline position fluctuation curve simulation method and device and electronic equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108769617A (en) * | 2018-06-25 | 2018-11-06 | 浙江大学 | Shipping depth gauge based on unmanned plane reads intelligent identifying system |
CN109711353A (en) * | 2018-12-28 | 2019-05-03 | 中国矿业大学 | A kind of drauht line area recognizing method based on machine vision |
CN109903303A (en) * | 2019-02-25 | 2019-06-18 | 秦皇岛燕大滨沅科技发展有限公司 | A kind of drauht line drawing method based on convolutional neural networks |
-
2020
- 2020-03-26 CN CN202010222648.9A patent/CN111476120B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108769617A (en) * | 2018-06-25 | 2018-11-06 | 浙江大学 | Shipping depth gauge based on unmanned plane reads intelligent identifying system |
CN109711353A (en) * | 2018-12-28 | 2019-05-03 | 中国矿业大学 | A kind of drauht line area recognizing method based on machine vision |
CN109903303A (en) * | 2019-02-25 | 2019-06-18 | 秦皇岛燕大滨沅科技发展有限公司 | A kind of drauht line drawing method based on convolutional neural networks |
Non-Patent Citations (1)
Title |
---|
基于卷积神经网络的船舶水尺字符识别方法研究;杜佳峰等;《中国水运》;20200315;第1-3页 * |
Also Published As
Publication number | Publication date |
---|---|
CN111476120A (en) | 2020-07-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111476120B (en) | Unmanned aerial vehicle intelligent ship water gauge identification method and device | |
CN111626190B (en) | Water level monitoring method for scale recognition based on clustering partition | |
CN110378308B (en) | Improved port SAR image near-shore ship detection method based on fast R-CNN | |
CN109740460B (en) | Optical remote sensing image ship detection method based on depth residual error dense network | |
CN110276285B (en) | Intelligent ship water gauge identification method in uncontrolled scene video | |
CN111652149A (en) | Deep convolutional neural network-based benthonic oil sonar detection image identification method | |
CN109427055B (en) | Remote sensing image sea surface ship detection method based on visual attention mechanism and information entropy | |
CN111476159B (en) | Method and device for training and detecting detection model based on double-angle regression | |
CN111626993A (en) | Image automatic detection counting method and system based on embedded FEFnet network | |
CN108509919B (en) | Method for detecting and identifying waterline in video or picture based on deep learning | |
CN101915910B (en) | Method and system for identifying marine oil spill object by marine radar | |
CN110619328A (en) | Intelligent ship water gauge reading identification method based on image processing and deep learning | |
CN109711353B (en) | Ship waterline area identification method based on machine vision | |
CN109064479B (en) | Sea-sky-line detection method based on gray dynamic features of adjacent video frames | |
CN105741276A (en) | Ship waterline extraction method | |
CN111723632B (en) | Ship tracking method and system based on twin network | |
CN116665095B (en) | Method and system for detecting motion ship, storage medium and electronic equipment | |
CN116152115B (en) | Garbage image denoising processing method based on computer vision | |
CN104021553A (en) | Sonar image object detection method based on pixel point layering | |
Yang et al. | Evaluation and mitigation of rain effect on wave direction and period estimation from X-band marine radar images | |
CN115393734A (en) | SAR image ship contour extraction method based on fast R-CNN and CV model combined method | |
CN114764801A (en) | Weak and small ship target fusion detection method and device based on multi-vision significant features | |
CN116434230A (en) | Ship water gauge reading method under complex environment | |
CN117197048B (en) | Ship water gauge reading detection method, device and equipment | |
CN113344148A (en) | Marine ship target identification method based on deep learning |
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 | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20231114 Address after: Building 78, Yard 12, Zhuyuan Road, Shunyi District, Beijing, 101300 (Tianzhu Comprehensive Bonded Zone) Patentee after: Lihong Inspection Group Co.,Ltd. Address before: 434023 No. 1 South Ring Road, Hubei, Jingzhou Patentee before: Yangtze University |