CN111339968A - Method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish based on computer vision technology - Google Patents
Method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish based on computer vision technology Download PDFInfo
- Publication number
- CN111339968A CN111339968A CN202010134302.3A CN202010134302A CN111339968A CN 111339968 A CN111339968 A CN 111339968A CN 202010134302 A CN202010134302 A CN 202010134302A CN 111339968 A CN111339968 A CN 111339968A
- Authority
- CN
- China
- Prior art keywords
- fish
- image
- acquiring
- fish school
- computer vision
- 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.)
- Withdrawn
Links
- 241000251468 Actinopterygii Species 0.000 title claims abstract description 85
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000009360 aquaculture Methods 0.000 title claims abstract description 36
- 244000144974 aquaculture Species 0.000 title claims abstract description 36
- 238000005516 engineering process Methods 0.000 title claims abstract description 15
- 230000009466 transformation Effects 0.000 claims abstract description 13
- 230000000877 morphologic effect Effects 0.000 claims abstract description 11
- 238000004364 calculation method Methods 0.000 claims abstract description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 23
- 206010000117 Abnormal behaviour Diseases 0.000 claims description 6
- 230000002708 enhancing effect Effects 0.000 claims description 6
- 238000002834 transmittance Methods 0.000 abstract description 6
- 230000002159 abnormal effect Effects 0.000 description 5
- 239000013505 freshwater Substances 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000007547 defect Effects 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 229910052760 oxygen Inorganic materials 0.000 description 2
- 239000001301 oxygen Substances 0.000 description 2
- 238000002310 reflectometry Methods 0.000 description 2
- 230000009182 swimming Effects 0.000 description 2
- 238000000889 atomisation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/35—Categorising the entire scene, e.g. birthday party or wedding scene
- G06V20/36—Indoor scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- 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
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a computer vision technology-based two-dimensional coordinate acquisition method for indoor high-density aquaculture fish, which comprises the following steps: acquiring a cultured fish image; optimizing an atmospheric scattering model; acquiring a dark channel image; optimizing the transmittance; morphological treatment; dividing a fish target in a watershed; and calculating the two-dimensional coordinates of the fish target. The invention improves the prior theoretical algorithm of the dark channel originally used for defogging the outdoor image, not only optimizes the transmissivity calculation method, but also can adjust the atmospheric light intensity value according to the actual condition of the image. The processed fish school image has rich individual detail information, strong layering sense and strong contrast. On the basis, distance transformation and morphological reconstruction are carried out on the constructed binary image, and a watershed algorithm is used for effectively segmenting the adhered individuals in the fish school, so that guarantee is provided for reliably calculating the physical center coordinates of all the bodies of the fish school.
Description
Technical Field
The invention belongs to the technical field of intelligent aquaculture, and particularly relates to a computer vision technology-based method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish.
Background
The fresh water aquaculture area of China accounts for 73.04% of the total area, and the fresh water aquaculture area has an important position in the national economic system of China for a long time. The water resources of the regions in the middle and downstream of the Yangtze river are rich, and the method is a main region for freshwater aquaculture in China. Influenced by subtropical monsoon climate, the areas in the middle and lower reaches of Yangtze river generally have high temperature, high pressure and much rain in summer, and belong to the high-risk period of freshwater aquaculture. Only in the area, the shortage of dissolved oxygen in the water body is not timely remedied every year, and the loss of aquaculture reaches hundreds of millions of yuan.
The water quality sensor is used for measuring the water quality in real time, and the water quality sensor is inconsistent with the current national conditions of China due to the defects of high price, short service life, difficult maintenance and the like. Aiming at indoor high-density aquaculture, the water surface two-dimensional coordinates of the fish school in the aquaculture pond are obtained, and the abnormal behavior of the fish school is analyzed according to characteristic parameters such as the swimming speed, the cluster characteristics, the floating head behavior and the like of the aquaculture fish on the basis, so that the aquaculture water quality is timely and reliably warned. The method has great significance for freshwater aquaculture in China and has great reference value for water environment protection.
By means of a two-dimensional computer image data acquisition system, the water surface fish image of the indoor aquaculture pond is acquired, and the accurate acquisition of two-dimensional centroid coordinates of each body of the fish is influenced because the fish image acquired from the water surface has the defects of blurring, atomization and the like.
The atmospheric scattering model is simplified, and the dark channel prior theory algorithm originally used for outdoor image defogging is further optimized and improved, so that the indoor fish school image is defogged and enhanced, the atmospheric light intensity value can be adaptively adjusted according to the actual condition of the image, and the transmissivity calculation method is optimized. The method can effectively improve the definition of each body in the fish school image, so that the fish school individual detail information is rich, the layering sense is strong, and the contrast and the visual effect are strong. On the basis, the constructed binary image is subjected to morphological denoising treatment, the image is subjected to distance transformation and morphological reconstruction, the adhesion part of the fish school in the image is effectively segmented by using an optimized watershed algorithm, and the two-dimensional centroid coordinates of each individual target in the fish school are finally determined by using a method for determining the centroid of each region. Through the knowledge of the coordinate change of the two-dimensional mass center of each body in the fish school, the behavior of the cultured fish under the condition of abnormal water quality is obtained to a certain extent, and the method has important significance for improving the intelligent level of indoor high-density aquaculture.
Disclosure of Invention
The invention relates to a method for acquiring two-dimensional coordinates of indoor aquaculture fish based on a computer vision technology, aiming at the technical problem of the existing indoor high-degree industrial aquaculture, and the method is used for analyzing the behavior characteristics of the aquaculture fish under the condition of abnormal water quality on the basis and finally applying the behavior characteristics to the early warning of the abnormal aquaculture water quality, so that the economic risk of farmers is reduced to the greatest extent, the intelligent management level of the indoor industrial aquaculture is enhanced, and the problems in the prior art are solved.
In order to achieve the purpose, the invention provides the following technical scheme: a two-dimensional coordinate acquisition method for indoor high-density aquaculture fish based on a computer vision technology comprises the following steps:
step 1, acquiring a fish school image;
step 2, enhancing the fish school image;
step 3, removing adhesion of the fish school target and dividing;
step 4, calculating two-dimensional coordinates of the fish target;
and 5, characterizing abnormal behavior of the fish school.
Compared with the prior art, the invention provides a computer vision technology-based method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish, which has the following beneficial effects:
the invention can realize the accurate acquisition of two-dimensional coordinates of the indoor high-degree industrial aquaculture fish school on the water surface of the culture pond by means of the computer vision technology and the optimized image processing algorithm, analyzes and obtains the behavior characteristics of the cultured fish under the condition of abnormal water quality on the basis, provides scientific basis for the accurate prediction of the cultured water quality, and greatly reduces the economic risk of farmers.
Drawings
FIG. 1 is a flow diagram of the present invention.
Detailed Description
The invention provides a computer vision technology-based method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish on the water surface of an aquaculture pond, aims to better and scientifically early warn aquaculture water quality in time and reliably based on abnormal behavior characteristics of fish schools, and accordingly greatly reduces aquaculture risks and solves the problems in the prior art.
In order to achieve the purpose, the invention provides the following technical scheme: a fish school two-dimensional coordinate obtaining method applied to indoor high-density aquaculture comprises the step of obtaining images of cultured fish on the water surface by using an indoor culture pond observation camera. Aiming at image defects, an atmospheric scattering model is simplified, and defogging and enhancement processing are carried out on an image by adopting an optimized dark channel prior theory algorithm. On the basis, morphological denoising processing is carried out on the enhanced fish school binary image, and on the basis of carrying out distance transformation and morphological reconstruction on the fish school image, effective segmentation is carried out on each adhered individual target in the fish school by adopting an optimized watershed algorithm. And finally, realizing the centroid coordinates of each individual target in the fish school by using a method for determining the centroid of each area. According to the fish school coordinate parameters, abnormal behavior characteristics such as swimming speed, cluster characteristics, floating head behavior and the like of the cultured fish can be obtained and are used as the basis for early warning of the abnormal culture water quality.
Referring to fig. 1, the present invention provides a technical solution: a two-dimensional coordinate acquisition method for indoor high-density aquaculture fish based on a computer vision technology comprises the following steps:
step 1, acquiring a fish school image;
step 2, enhancing the fish school image;
step 3, removing adhesion of the fish school target and dividing;
step 4, calculating two-dimensional coordinates of the fish target;
and 5, characterizing abnormal behavior of the fish school.
In this embodiment, in the step 1, in acquiring the fish school image, the method includes: and an indoor observation camera is used for acquiring the image of the fish cultured in the net cage on the water surface.
In this embodiment, the step 2, in enhancing the fish image, includes the following steps:
2.1, simplifying an atmospheric scattering model;
and 2.2, defogging and enhancing the image by adopting an optimized dark channel prior theory algorithm.
In this embodiment, in the step 3, in the fish school target adhesion removal and segmentation, the method includes: and carrying out morphological denoising on the enhanced fish school binary image, and effectively segmenting each adhered individual target in the fish school by adopting an optimized watershed algorithm on the basis of carrying out distance transformation and morphological reconstruction on the fish school image.
In this embodiment, in the step 4, in the calculation of the two-dimensional coordinate of the fish target, the method includes: and realizing the centroid coordinates of each individual object in the fish school by using a method for determining the centroid of each area.
The specific embodiment is as follows: generally, cameras are used in an environment where air media are uniform, and the atmospheric scattering model can be simplified as follows:
I(x)＝Aρ(x)t(x)+A(1-t(x)) (1)
wherein: i (x): the fog image actually taken by the camera;
ρ (x): standard scene reflectivity (normalized radiation of the photographed target);
a: sky (atmospheric) light intensity;
t (x): transmittance, the reflected light of the photographed object that effectively enters the camera;
the defogging enhancement processing on the image is based on the model of formula 1. In the invention, the calculation process of obtaining the atmospheric light intensity A from the dark channel prior theory by the He is optimized, and the value of the atmospheric light intensity A is approximately determined by adopting an interval estimation method. According to the dark channel prior theory, minimizing two sides of formula 1 to obtain:
in formula 2, j (x) is an image without haze. The three color channels of the image are processed simultaneously according to the dark channel prior theory Idark(x)＝A(1-t(x))(Idark(x) A dark channel containing the fog image I), then
Therefore, the atmospheric light intensity A is compared with the image Idark(x) Is larger than the maximum pixel value in the image i (x), and is smaller than the maximum pixel value in the image i (x), namely, the maximum pixel value in the image i (x) is satisfied
The atmospheric light intensity a can be estimated as:
wherein,
value is pair
All pixel values contained in (a) are averaged.
The experiment of acquiring the fish school behavior image under the condition of lack of dissolved oxygen is carried out indoors, so that the color in the image is deviated. Therefore, when the image is defogged, the image needs to be corrected to be white by using white balance processing (WP algorithm, also called MAX-RGB algorithm). Color deviations in the image can be eliminated well. Thereby further converting the atmosphere scattering model:
according to the dark channel prior theory, there is
The transmittance t' (x) in the atmospheric scattering model is shown as equation 5:
in evaluating the transmittance, the transmittance calculation needs to be optimized first. The invention can make up J in a certain degree of self-adaptability according to the actual fuzzy condition of the image by setting the function w (x)dark(x) Not equal to 0. Using the previously estimated atmospheric light intensity value A and the dark channel image I containing the fog imagedark(x)(Idark(x) < A), establishing a function
w(x)：
Then, the optimized transmittance t (x) is obtained as:
to this end, the image scene reflectivity can be obtained by the simplified atmospheric scattering model equation 4
From j (x) ═ ρ (x) a, a sharp image can be obtained.
The invention improves a watershed algorithm, and provides a simple and efficient method for segmenting each body in a fish school image based on watershed of distance transformation.
"distance transform" of a binary image refers to the distance from each pixel to the pixel closest to the zero value. The distance here is "Euclidean distance transform"Method for the converted binary image Jm×n＝[aij]Divided into "background pixels" (set B { (x, y) | a)ij0) and "target pixel" (set C { (x, y) | a)ij1}) two. But the binary image cannot reflect some important feature information in the image. By "distance transformation", the pixel value of each pixel in the binary image is converted into a minimum distance value between the pixel and all background pixels in the image. In this way, pixels which do not have any relationship with each other in the original binary image can be converted into a gray-scale image which can be displayed as an image skeleton, an edge and a relative position between the pixels. For binary image Jm×nThe distance transformation performed on all the pixels (i, j) in the equation 8 is as follows:
in order to avoid 'over-segmentation' of the fish school, the invention further modifies the transformation result after the distance transformation is carried out on the image so as to avoid an undesirable 'local minimum' region. The specific method is that morphological reconstruction is carried out on the fish school image after distance transformation. First, for low-brightness regions in the image, regions in the vicinity of the "local minimum" in the image that differ by a certain threshold (2 in this example) are determined, i.e., small dots are generated only in the regions that are desired to be segmented in the image. And then the image is superposed with the original binary image, some undesired local minimum areas which appear after the distance transformation are filtered through the operation of 'forced minimum', and the modification of the result after the distance transformation is finished, wherein the modification result is as follows: the "local minimum" is retained only at the location where segmentation is desired. The method effectively avoids the over-segmentation problem caused by the fact that the watershed algorithm is used for segmenting the fish school later. And finally, after the segmented fish school binary image is converted into a 'labeling matrix', the two-dimensional coordinates of each body in the fish school on the water surface of the culture pond are determined by a method for determining the mass center of each region.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can substitute or change the technical solution of the present invention and the inventive concept within the technical scope of the present invention, and the technical solution and the inventive concept thereof should be covered by the scope of the present invention.
Claims (5)
1. A two-dimensional coordinate acquisition method for indoor high-density aquaculture fish based on computer vision technology is characterized by comprising the following steps:
step 1, acquiring a fish school image;
step 2, enhancing the fish school image;
step 3, removing adhesion of the fish school target and dividing;
step 4, calculating two-dimensional coordinates of the fish target;
and 5, characterizing abnormal behavior of the fish school.
2. The method for acquiring the two-dimensional coordinates of the indoor high-density aquaculture fish based on the computer vision technology, as claimed in claim 1, is characterized in that: in the step 1, in acquiring the fish school image, the method comprises the following steps: and an indoor observation camera is used for acquiring the image of the fish cultured in the net cage on the water surface.
3. The method for acquiring the two-dimensional coordinates of the indoor high-density aquaculture fish based on the computer vision technology, as claimed in claim 1, is characterized in that: in the step 2, the fish school image enhancement comprises the following steps:
2.1, simplifying an atmospheric scattering model;
and 2.2, defogging and enhancing the image by adopting an optimized dark channel prior theory algorithm.
4. The method for acquiring the two-dimensional coordinates of the indoor high-density aquaculture fish based on the computer vision technology, as claimed in claim 1, is characterized in that: in the step 3, in the adhesion removing and dividing of the fish school target, the method comprises the following steps: and carrying out morphological denoising on the enhanced fish school binary image, and effectively segmenting each adhered individual target in the fish school by adopting an optimized watershed algorithm on the basis of carrying out distance transformation and morphological reconstruction on the fish school image.
5. The method for acquiring the two-dimensional coordinates of the indoor high-density aquaculture fish based on the computer vision technology, as claimed in claim 1, is characterized in that: in the step 4, in the calculation of the two-dimensional coordinate of the fish target, the method comprises the following steps: and realizing the centroid coordinates of each individual object in the fish school by using a method for determining the centroid of each area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010134302.3A CN111339968A (en) | 2020-02-28 | 2020-02-28 | Method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish based on computer vision technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010134302.3A CN111339968A (en) | 2020-02-28 | 2020-02-28 | Method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish based on computer vision technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111339968A true CN111339968A (en) | 2020-06-26 |
Family
ID=71185865
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010134302.3A Withdrawn CN111339968A (en) | 2020-02-28 | 2020-02-28 | Method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish based on computer vision technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111339968A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111783745A (en) * | 2020-08-06 | 2020-10-16 | 珠海南方利洋水产科技有限公司 | Fish health judgment method and device applied to pond culture and computer-readable storage medium |
-
2020
- 2020-02-28 CN CN202010134302.3A patent/CN111339968A/en not_active Withdrawn
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111783745A (en) * | 2020-08-06 | 2020-10-16 | 珠海南方利洋水产科技有限公司 | Fish health judgment method and device applied to pond culture and computer-readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106548463B (en) | Sea fog image automatic defogging method and system based on dark and Retinex | |
| CN106530246B (en) | Image defogging method and system based on dark Yu non local priori | |
| CN108596849B (en) | Single image defogging method based on sky region segmentation | |
| CN107301623B (en) | Traffic image defogging method and system based on dark channel and image segmentation | |
| Le et al. | An automated fish counting algorithm in aquaculture based on image processing | |
| CN107767354A (en) | A kind of image defogging algorithm based on dark primary priori | |
| Li et al. | Detection of uneaten fish food pellets in underwater images for aquaculture | |
| CN110211067B (en) | Defogging method for UUV visible light image on offshore surface | |
| CN109858480A (en) | Digital instrument identification method | |
| CN116091421A (en) | Method for automatically dividing and calculating area of blastomere image of in-vitro fertilized embryo | |
| CN109270853A (en) | A kind of working method of intelligent ecological Water Management System | |
| CN109272475B (en) | Method for rapidly and effectively repairing and strengthening underwater image color | |
| CN105913425B (en) | A kind of more pig contour extraction methods based on adaptive oval piecemeal and wavelet transformation | |
| CN105989583B (en) | A kind of image defogging method | |
| CN113516071B (en) | Pig weight estimation method | |
| CN103177244B (en) | Method for quickly detecting target organisms in underwater microscopic images | |
| CN111339968A (en) | Method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish based on computer vision technology | |
| CN112907461B (en) | Defogging enhancement method for infrared foggy-day degraded image | |
| CN113327263B (en) | Image vision-based fish school activity monitoring method | |
| CN116843581B (en) | Image enhancement method, system, device and storage medium for multi-scene graph | |
| CN113763404A (en) | Foam image segmentation method based on optimization mark and edge constraint watershed algorithm | |
| CN110533619A (en) | A kind of image defogging improved method based on dark channel prior | |
| CN116229404A (en) | Image defogging optimization method based on distance sensor | |
| CN113012167B (en) | Combined segmentation method for cell nucleus and cytoplasm | |
| CN108830834A (en) | A kind of cable-climbing robot video artefacts information automation extraction method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20200626 |
|
| WW01 | Invention patent application withdrawn after publication |