JP4411576B2 - Aquatic animal counting equipment - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an aquatic animal counting processing apparatus that counts the number of aquatic animals cultivated using a breeding board.
[0002]
[Prior art]
Growing plates are used for aquaculture of aquatic animals such as sea urchins and abalone. Then, when managing the growth situation of aquatic animals during aquaculture, the number of aquatic animals attached to the breeding board or the number of individuals by size is counted. Such counting of the number of aquatic animals is conventionally performed visually.
[0003]
[Problems to be solved by the invention]
Conventionally, the number of individuals of aquatic animals is counted visually. The counting of the number of individuals is performed on several tens to several hundreds of growth plates at a time, and several hundreds of aquatic animals are attached to one growth plate. Therefore, the labor load required for the counting work is increased. In addition, if algae such as diatoms adhere to the growth plate, it is difficult to distinguish between aquatic animals and algae, which increases work time and labor load and causes erroneous counting. As a result, there arises a problem that an aquatic animal shipment plan cannot be properly performed.
[0004]
An object of the present invention is to provide an aquatic animal counting processing apparatus that solves the above-described drawbacks and facilitates the counting operation of the number of aquatic animals adhering to a growth plate.
[0005]
[Means for Solving the Problems]
In the aquatic animal counting processing apparatus for counting the number of aquatic animals cultivated on a cultivating plate, the present invention provides a storage means for storing the cultivating plate and a light for the cultivating plate stored in the storing means. A light projecting unit that irradiates; an image capturing unit that captures the transmitted light that has passed through the growth plate; converts the captured transmitted light into a transmitted light image; and an image processing unit that processes the transmitted light image. The processing means has a first counting means for counting the number of individuals of the aquatic animals.
[0006]
DETAILED DESCRIPTION OF THE INVENTION
An embodiment of the present invention will be described with reference to FIG. The storage means 11 is a storage container configured to prevent light from the outside from entering, for example, a light shielding box in which each peripheral wall portion is formed of a light shielding member, and a growth plate 12 to which aquatic animals are attached is housed. Is done. The growth plate 12 is a thin plate having a corrugated cross section, for example, and guides 13 a and 13 b are provided above and below the growth plate 12. A moving means 14, such as a direct acting actuator, is fixed to a part of the guides 13a and 13b. The moving means 14 is connected to the growing plate 12 via the connecting rod 15, and when the moving means 14 enters the operating state, the growing plate 12 moves in the lateral direction indicated by the arrow Y along the guides 13a and 13b.
[0007]
A light projecting means 16 is disposed on one side of the growth plate 12. The light projecting means 16 is entirely formed in a waterproof structure, and is composed of, for example, a light source or a light diffusion plate in which a plurality of light emitting diodes (hereinafter referred to as LEDs) are arranged on a substrate. It comes to irradiate.
[0008]
On the other side of the growth plate 12, an image pickup means 17, for example, twelve CCD cameras 1701 to 1712 are arranged in the vertical direction. The imaging means 17 images the transmitted light emitted from the light projecting means 16 and transmitted through the growth plate 12, and converts the captured transmitted light into, for example, an electrically transmitted light image and outputs it. Each of the CCD cameras 1701 to 1712 is connected to the switcher 18. The switch 18 is sequentially switched in connection state under the control of the controller 19, and transmitted light images of the CCD cameras 1701 to 1712 in the connection state are sent to the image processing means 20.
[0009]
The growing plate 12 has a corrugated cross section, but a flat growing plate can also be used. Further, the growth plate 12 is divided into, for example, 12 pieces in the vertical direction, divided into, for example, 5 pieces in the horizontal direction, and the entire region is divided into, for example, 60 divisions. Then, the transmitted light in 12 sections located in the extreme end in the horizontal direction, for example, in the vertical direction, is captured by the CCD cameras 1701 to 1712, converted into an electrical transmitted light image, and output. These transmitted light images are sequentially sent to the image processing means 20 for each of the CCD cameras 1701 to 1712 by switching the switch 18.
[0010]
Thereafter, the moving means 14 operates, and the growth plate 12 moves by one section in the lateral direction and stops. At that position, 12 sections of transmitted light transmitted through the growth plate 12 are imaged by the CCD cameras 1701 to 1712 and converted into an electrically transmitted light image. Such operations are repeated in order, and the entire region of the growth plate 12, for example, 60 sections of transmitted light images are output from the imaging unit 17 and supplied to the image processing unit 20.
[0011]
The image processing means 20 includes an image processing unit 201, a counting unit 202, a recording display unit 203, and the like. The image processing unit 201 performs various types of image processing on the transmitted light image sent from the imaging unit 17, removes unnecessary transmitted light images included in the transmitted light image, for example, transmitted light images such as algae, and the like. Extract the animal's transmitted light image. The counting unit 202 counts the number of individuals of aquatic animals extracted by the image processing unit 201 and the number of individuals by size. The recording display unit 203 records and displays the counting result of the counting unit 202 and outputs the counting result as an electric signal or the like to an external device.
[0012]
Here, the relationship between the light emitted from the light projecting means 16 and the transmitted light transmitted through the growth plate 12 will be described with reference to FIG. The horizontal axis in FIG. 2 is the wavelength of light (nm), the vertical axis is the attenuation (dB), and the curve D shows the measurement result of the spectral intensity distribution of the transmitted light that has passed through the algae. As can be seen from FIG. 2, algae have a characteristic of transmitting light having a wavelength of about 800 nm to 900 nm or more. Therefore, for example, an infrared LED having a wavelength of 945 nm is used as the light source of the light projecting means 16. Note that, when a light source is selected, light including a wavelength of 800 nm to 900 nm or more, particularly 900 nm or more is suitable, although it depends on the frequency sensitivity characteristics of the imaging means.
[0013]
Next, a transmitted light image of the growth plate to which algae is attached will be described with reference to FIG. FIG. 3A shows a case where an infrared LED (wavelength 945 nm) is used as a light source, and FIG. 3B shows a case where a white LED is used as a light source. When a white LED is used as shown in FIG. 3B, it is difficult to separate the aquatic animals W1 to W3 indicated by diagonal lines from the algae X indicated by dots. As shown in FIG. 3A, when an infrared LED is used, light is almost transmitted through the algae X, and the shadows of the aquatic animals W1 to W3 in the algae X are clear. Therefore, even when the aquatic animal attached to the growing plate is located on or in the algae, these unnecessary backgrounds and aquatic animals are separated, and the counting accuracy is improved.
[0014]
Next, image processing performed by the image processing means 20, a counting method, and the like will be described with reference to the flowcharts of FIGS.
[0015]
First, image correction is performed on the transmitted light image sent from the imaging means 17 (FIG. 1) (S1). The transmitted light image of the entire area of the growth plate is divided into, for example, 12 sections in the vertical direction and 5 sections in the horizontal direction, and is divided into 60 sections as a whole. In this case, one section corresponds to a transmitted light image output from one CCD camera, for example, and is composed of, for example, 640 pixels × 480 pixels. The transmitted light image is divided into 256 hierarchies, for example, from black level 0 to white level 255.
[0016]
In general, aquatic animals do not transmit light, and result in a black transmitted light image. Therefore, in the image correction (S1), for example, an image of a bright region having a hierarchy of 200 or more is determined as a background that does not include aquatic animals, and these transmitted light images are removed.
[0017]
In this way, by removing bright image portions having a gray level distribution of 200 or more at the initial stage of image correction processing, the number of data to be subjected to subsequent image processing is reduced and image processing is speeded up.
[0018]
Next, a transmitted light image of one section is cut out from a predetermined area of the growing plate 12, for example, the entire area of the growing plate 12 (S2), and the distribution of the transmitted light image in the one section, for example, as shown in FIG. A histogram is created (S3). The horizontal axis in FIG. 7 indicates the gray level, and the vertical axis indicates, for example, the number of pixels. In the following description, the level L where the transmitted light image close to black with the highest density appears on the histogram is the minimum position, the level P where the most transmitted light image appears is the peak position, and the transmitted light image close to white with the lowest density. The hierarchy M in which “” appears is called the maximum position. Based on this histogram, for example, the binarized effective range E including the aquatic animals, that is, the reference value is set from the minimum position to (peak position−10th hierarchy).
[0019]
Next, one section is further divided into, for example, a plurality of subsections (20 pixels × 20 pixels) (S4), and the transmitted light image in the one subsection is 2 based on the reference value set in S3. The image is converted into a binary image (S5).
[0020]
Next, a lump (hereinafter referred to as a graphic) in which one or a plurality of pixels of the binarized image are collected is numbered and so-called labeling is performed (S6).
[0021]
Next, for example, using the data used when creating the density distribution in S3, a histogram of the transmitted light image is created for the labeled figure, and the number of gray levels from the minimum position to the peak position is a predetermined reference value. For example, only data within 20 is extracted as valid data (S7). Aquatic animals such as sea urchins have a characteristic of rising steeply with a small number of layers from the minimum position to the peak position, while algae have a characteristic of gradually rising with a large number of layers from the minimum position to the peak position. Therefore, the transmitted light image of the algae is deleted by the process of S7, and the transmitted light image of the aquatic animal is extracted.
[0022]
Next, the image of the small section subjected to the above-described processing of S4 to S7 is returned to the original location of one section (S8). Thereafter, the processes of S4 to S8 are sequentially repeated for the other small-section images, and the transmitted light images of all the sections are processed.
[0023]
When the image processing for one section is completed, a figure with a small area, for example, a figure having a pixel number of 10 or less is removed as a particle lump for a figure labeled as valid data in the one section. (S9).
[0024]
Next, span removal is performed on the graphic labeled as valid data in the process of S9 (S10). In span removal, a histogram of density distribution is created for each labeled figure, and a figure whose span width from the minimum position to the maximum position is a predetermined value, for example, 200 layers or more, is removed. In this case, since the aquatic animal has a short span width and one algae has a long span width, only the figure of the aquatic animal is effectively extracted by deleting the figure having the large span width.
[0025]
Next, for a figure extracted in the process of S10, a figure having a histogram determination rate of, for example, 50% or less is removed (S11).
[0026]
In this case, the fixed rate is, for example, fixed rate = {1 one (peak position−minimum position) / effective width)} (1)
It is defined as In the following description, “%” obtained by multiplying the expression (1) by 100 is used as a numerical value of the determination rate.
[0027]
For example, an aquatic animal has a characteristic of rapidly rising from a minimum position to a peak position in a layer with little shading, and a definite rate (%) increases. Algae and the like have a characteristic of gradually rising from the minimum position to the peak position, so the determination rate (%) becomes small. Therefore, a figure of aquatic animals is extracted by removing a figure having a deterministic rate of, for example, 50% or less.
[0028]
Next, the graphic data deleted by the span removal in S10 is recalled once again, and a graphic whose span width of the histogram is a specified value, for example, (20th to 160th) is taken out as valid data and binarized. Then, span removal is performed on the figure of the binarized image (S12). In this case, a span width narrower than the span removal of S10 is set as a reference value, and, for example, within 50 layers is extracted as valid data. The data deleted in the first span removal is stored in the buffer, and the data stored in the buffer is recalled and used in the second span removal.
[0029]
Next, among the labeled figures, a figure with a small area, for example, 10 pixels or less is removed as a particle lump (S13).
[0030]
And also about the transmitted light image of another division, the process of said S9-S13 is repeated in order, the transmitted light image of all the divisions is processed, for example, is synthesize | combined as a figure of the whole area | region of a growth board.
[0031]
Next, the distance from the background to each pixel is calculated for the graphic determined to be valid data over the entire area of the growth plate (S14). In this case, the distance between the background and the pixel is the square root of the sum of (the square of the number of vertical pixels) and (the square of the number of horizontal pixels) from the nearest background pixel such as a white pixel to each pixel constituting the figure. Calculated.
[0032]
Next, for example, 30% or more of data is validated and graphic separation is performed on the labeled graphic in the entire area of the growth plate (S15).
[0033]
Next, the distance between the background and the pixels is calculated again for the graphics in the entire area of the growth plate that has been separated (S16).
[0034]
Next, a numerical value larger than S15, for example, 60% or more data is validated and graphic separation is performed (S17).
[0035]
In the graphic separation process described above, graphic separation is performed twice using different reference values. The reason will be described with reference to FIG. The code | symbol W of FIG. 8 has shown the figure of the aquatic animal, and the code | symbol X has shown the figure of the algae.
[0036]
FIG. 8A shows a figure in which the overlapping state of objects such as aquatic animals W and algae X is large and the contour pattern is less constricted, and FIG. 8B is a figure in which the overlapping state of the object is small and the contour pattern is narrowed Z. The figure is shown. If the constriction Z is clear, if the figure separation is validated with a large value of 60% or more from the beginning, the periphery of the figure may be deleted and no valid data may remain. Therefore, the figure separation is performed a plurality of times, for example, twice.
[0037]
Next, the pattern of graphic separation will be described with reference to FIG.
[0038]
FIG. 9A shows an example of a state in which the distance from the background to each pixel of the figure is calculated, and the numbers indicate the distance from the background (white area). Here, for example, a case where there are three sea urchins having distances of “6”, “8”, and “4” is shown as an example, and three sea urchins are included in one graphic G. The numerical value indicating the distance is a number selected for convenience of explanation, and some of the numbers do not necessarily correspond to the distance from the background (white area).
[0039]
FIG. 9B is a diagram in which 30% or more of the peak “8” in FIG. 9A is validated, and “6” “8” “4” “3” of 8 × 0.3 = 2.4 or more. ”Data is cut out by being validated and separated into two figures G1 and G2.
[0040]
FIG. 9C shows a state in which the distance is calculated again for the cut out data. In this case, the peak is “6”, and when 60% is validated, data of 6 × 0.6 = 3.6 or more is regarded as valid, and the portions “4” and “6” are represented by G11 and G12. It is cut out, and finally, as shown in FIG. 9 (d), it is cut into three figures G11, G12, and G2, and it is determined that there are three.
[0041]
When the graphic separation process described above is used, even when aquatic animals are touching or overlapping, the number and size of the aquatic animals can be correctly counted, and the counting accuracy is improved.
[0042]
Next, after graphic separation, among the labeled figures, a figure with a small area, for example, 10 pixels or less is removed as a particle lump (S18).
[0043]
Next, the figure cut out as valid data in each of the above processes is labeled (S19), and the number of the labeled figures is counted by the counting unit 202 (FIG. 1). The counting unit 202 measures at least one of the figure size, for example, the number of pixels and the diameter (maximum width), and the number of figures (number of aquatic animals) or the number of figures by size (number of aquatic animals). Counted (S20).
[0044]
Next, the counting result of S20 is displayed and recorded on the recording display unit 203 (FIG. 1) (S21).
[0045]
Next, it is determined whether to continue the counting process (S22). If YES, the process returns to S1, and if NO, the process ends.
[0046]
In the above embodiment, when image processing is performed, the transmitted light image of the entire growth plate is divided into a plurality of sections, and one section is further divided into a plurality of small sections, for example, divided into three stages of large, medium, and small. . A part of the processing is performed in units of growth plates, a part of the processing is performed in units of sections, and a part of the processing is performed in units of small sections. However, all of the above image processing can be performed in one unit, for example, a growth plate unit, a division unit, or a small division unit. In addition, the image processing can be divided into two large and small units, and a part of the above-described image processing can be performed in a large unit, and the other part can be performed in a small unit.
[0047]
In the above embodiment, when the transmitted light image is binarized, the reference value is set in units of sections. If a common reference value is set for the entire growth plate, the density of the transmitted light image varies depending on the location of the growth plate, depending on the state of algae attachment, and the counting accuracy may be reduced. When the reference value is set in units of sections, a decrease in counting accuracy due to such variations is prevented. Further, by binarizing, even if the transmission state of aquatic animals and algae is not uniform and there is variation in the density of the transmitted light image, the data of the aquatic animal can be selected reliably.
[0048]
In the above embodiment, the entire growth plate is divided into a plurality of sections, and each section is imaged by the respective CCD camera. In this case, in order to improve the counting accuracy and facilitate image processing, it is desirable to associate one section with one CCD camera. Note that the number of CCD cameras constituting the image pickup means is not limited to 12, but one or any number can be used depending on conditions such as the size of the growth plate and the required resolution.
[0049]
In the above embodiment, the growth plate is moved. However, the imaging means can be moved. When the growth plate or the imaging means is moved, the range that one CCD camera can capture at a time can be narrowed, and the resolution is relatively improved. For this reason, it is not necessary to use imaging with a high-resolution camera or imaging by moving a line sensor, and a low-cost CCD camera can be used, thereby reducing costs.
[0050]
The numerical values used in the above-described embodiments are examples, and these numerical values are different depending on illumination conditions, imaging environments, and the like. Therefore, it is desirable to perform a preliminary test or the like and make a determination based on the preliminary test. The order of image processing is also an example, and the order can be changed as appropriate. Further, depending on the environment and conditions for counting aquatic animals, a part of the above processing can be omitted.
[0051]
According to the above configuration, aquatic animals inhabit the front and back sides of the breeding board, or a plurality of aquatic animals overlap, and even if the situation of diatoms and algae is not uniform, The number of aquatic animals contained in diatoms and algae can be accurately counted. In addition, the labor load and miscounts of workers are reduced. It is easy to automate the individual counting process, and as a result, a stable shipment plan is maintained.
[0052]
【The invention's effect】
ADVANTAGE OF THE INVENTION According to this invention, the marvelous animal count processing apparatus which can count easily the aquatic animal cultured with the breeding board is implement | achieved.
[Brief description of the drawings]
FIG. 1 is a configuration diagram for explaining an embodiment of the present invention.
FIG. 2 is a diagram for explaining an embodiment of the present invention and showing an example of a light intensity spectrum transmitted through algae.
FIG. 3 is a diagram for explaining an embodiment of the present invention, and is a diagram showing an image of light transmitted through a growth plate to which algae is attached.
FIG. 4 is a diagram showing a part of a flow for explaining an embodiment of the present invention.
FIG. 5 is a diagram showing a part of a flow for explaining an embodiment of the present invention.
FIG. 6 is a diagram showing a part of a flow for explaining an embodiment of the present invention.
FIG. 7 is a diagram for explaining an embodiment of the present invention and is a diagram showing an example of a histogram of a transmitted light image transmitted through a growth plate.
FIG. 8 is a diagram illustrating connection / separation of figures used in the embodiment of the present invention.
FIG. 9 is a schematic diagram for explaining connection and separation of figures used in the embodiment of the present invention.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 11 ... Storage means 12 ... Growth board 13a, 13b ... Guide 14 ... Moving means 15 ... Connecting rod 16 ... Light projection means 17 ... Imaging means 1701-1712 ... CCD camera 18 ... Switch 19 ... Controller 20 ... Image processing apparatus 201 ... Image processing 202 ... counting unit 203 ... recording display unit

Claims (7)

In the aquatic animal counting processing apparatus for counting the number of aquatic animals cultivated on a cultivating plate, storage means for storing the cultivating plate, and light projection for irradiating light to the cultivating plate stored in the storing means Imaging means for imaging the transmitted light transmitted through the growth plate, converting the captured transmitted light into a transmitted light image, and image processing means for processing the transmitted light image. Reference value calculation means for calculating a reference value from the minimum position, maximum position, and peak position of the gray level distribution of the transmitted light image converted by the imaging means, and the transmitted light image as a binarized image based on the reference value A histogram is created from the density distribution of the transmitted light image for each figure that constitutes the binarized image block and the binarized image block , and values calculated from the minimum position, maximum position, and peak position of the histogram are predetermined. compared to the value, below or Has a determination means for determining as valid data in the above case, and has a counting means for counting the number of individuals of the aquatic animals or counting the number of individuals according to the size and size of the aquatic animals. Aquatic animal counting processing equipment. The image processing means includes a reference value calculating means for calculating a reference value from the density distribution of the transmitted light image converted by the imaging means, and a binary value for converting the transmitted light image into a binarized image based on the reference value. And from a minimum position where a transmitted light image having the highest density appears on the density distribution of the transmitted light image constituting the binarized image block to a peak position where the most transmitted light image appears on the density distribution The aquatic animal counting processing apparatus according to claim 1, further comprising: a determination unit that determines that the data is valid when the gray level is less than or equal to a predetermined value. The image processing means includes a reference value calculating means for calculating a reference value from the density distribution of the transmitted light image converted by the imaging means, and a binary value for converting the transmitted light image into a binarized image based on the reference value. And a hierarchical width from a minimum position at which a transmitted light image having the highest density appears on a density distribution of the transmitted light image constituting the binarized image block to a maximum position at which the transmitted light image having the lightest density appears. The aquatic animal count processing apparatus according to claim 1, further comprising: a determination unit that determines that the data is valid when the value is equal to or less than a predetermined value. The image processing means includes a reference value calculating means for calculating a reference value from the gray level distribution of the transmitted light image converted by the imaging means, and a binary value for converting the transmitted light image into a binarized signal based on the reference value. On the grayscale distribution of the transmitted light image constituting the binarized image block,
Determining rate = 1− (gradation level from the maximum position to the peak position) / density level from the maximum position to the minimum position (however, the maximum position is the level where the transmitted light image with the lightest density appears on the density distribution, peak The position is the hierarchy where the most transmitted light image appears on the gray distribution, and the minimum position is the hierarchy where the transmitted light image with the highest density appears on the gray distribution)
The aquatic animal counting processing apparatus according to claim 1, further comprising a determination unit that determines that the data is valid when the determination rate is equal to or greater than a predetermined value. The image processing means includes a reference value calculating means for calculating a reference value from the density distribution of the transmitted light image converted by the imaging means, and a binary value for converting the transmitted light image into a binarized image based on the reference value. A distance calculating means for calculating a distance from the background of the transmitted light image constituting the binarized image block, and a predetermined coefficient for the distance of the transmitted light image calculated by the distance calculating means. The aquatic animal counting processing apparatus according to claim 1, further comprising: a determination unit that performs multiplication and determines that the data is valid data when the result of the multiplication is a predetermined value or more. The imaging unit divides the transmitted light image of the growth plate into a plurality of sections and captures each of the sections, so that one or more camera devices having moving means for moving one of the growth plate and the camera device relative to the other or, by a plurality of camera devices arranged on the classification of the breeding plates by dividing the entire surface at once includes one camera unit, either comprising a plurality for capturing, the image processing means on the classification unit The aquatic animal counting processing apparatus according to claim 1, wherein   The aquatic animal counting process according to claim 1, wherein the light projecting unit has a light source that generates light having a wavelength region of 800 nm or more, and the imaging unit has a characteristic capable of capturing light including the wavelength region of the light projecting unit. apparatus.

Images