JP6946015B2 - Underwater organism detection device, underwater organism detection system and underwater organism detection method - Google Patents

Description

Embodiments of the present invention relate to an aquatic organism detection device, an aquatic organism detection system, and an aquatic organism detection method.

When assessing the ecosystems of rivers and swamp lakes, it is important to understand the abundance of algae and other aquatic organisms. There are the following methods for detecting organisms in water. First, a large amount of water is sampled from the river to be detected. Allow the sampled water to settle statically. Place a few drops of precipitate on a grooved glass slide and cover the surface with a cover glass. Microscopic observation identifies the species of organisms in the water and counts their abundance. However, this method has a problem that it takes a lot of time for preprocessing and counting work.

Japanese Unexamined Patent Publication No. 09-75918

An object to be solved by the present invention is to provide an aquatic organism detection device, an aquatic organism detection system, and an aquatic organism detection method capable of detecting an aquatic organism in a shorter time.

The underwater organism detection device of the embodiment includes an image input unit, a wavelength band processing unit, and a shape discrimination processing unit. Image input unit from the imaging device you image Taking the flow cell raw water is passed through obtains image data generated on the basis of the light on the signal strength for each of the plurality of wavelength bands from the raw. The wavelength band processing unit extracts a pixel region including an image of algae to be detected contained in the raw water from the image of the image data based on the transmittance of each of the plurality of wavelength bands in the image data. The shape discrimination processing unit discriminates the algae to be detected based on the shape of the pixel region extracted by the wavelength band processing unit.

It is a figure which shows the structural example of the underwater organism detection system 1 in this embodiment. It is a figure which shows the specific configuration example of the image pickup apparatus 30. It is a figure which shows the specific example of a flow cell 40. It is a figure which shows the photograph example of algae which occurs in raw water. It is a figure which shows the spectroscopic spectrum of "diatom", "microcystis", "Anabaena", "Oscillatoria" and "formidium". It is a figure which shows the structural example of the underwater organism detection apparatus 50. It is a flow figure explaining the operation of the underwater organism detection device 50. It is a figure which shows the specific example of the image generated by the aquatic organism detection system 1.

Hereinafter, the aquatic organism detection device, the aquatic organism detection system, and the aquatic organism detection method of the embodiment will be described with reference to the drawings.

(Summary)
First, the outline of the aquatic organism detection system of the present embodiment will be described.
FIG. 1 is a diagram showing a configuration example of the aquatic organism detection system 1 in the present embodiment. As shown in FIG. 1, the aquatic organism detection system 1 includes a microscope 10, a light source 20, an imaging device 30, a flow cell 40, and an aquatic organism detection device 50. The image pickup device 30 and the underwater organism detection device 50 are communicably connected via the network 60. The network 60 may be a wide area communication network such as the Internet, or a premises communication network such as a LAN. The image pickup device 30 and the underwater organism detection device 50 may be connected by a cable instead of the network 60.

The microscope 10 generates a magnified image of a substance passing through the flow cell 40 based on the light emitted from the light source 20.

The light source 20 emits light in the direction in which the lens of the microscope 10 is located. The light source 20 may be configured by using any light emitting element. It is desirable that the light source 20 is configured by using a light emitting element whose wavelength peak is not characteristic. For example, it is desirable that a light emitting element other than a Light Emitting Diode (LED) is used for the light source 20. For example, the light source 20 may be configured by using a halogen lamp.

The image pickup apparatus 30 captures an image magnified by the microscope 10. The image pickup apparatus 30 has a spectroscopic function. The image pickup apparatus 30 captures, for example, an image of light having a plurality of different wavelengths for each wavelength. The image pickup device 30 transmits the image data obtained by the image pickup to the underwater organism detection device 50 via the network 60. The image pickup apparatus 30 may be configured by using, for example, a hyperspectral camera, or may be configured by using a color filter and a monochrome camera. Hereinafter, a specific configuration example of the image pickup apparatus 30 will be described.

First, a configuration when the imaging device 30 is configured by using a hyperspectral camera will be described. The image pickup device 30 includes an image pickup device in which pixels having a sensitivity capable of discriminating peaks of signal intensities in a plurality of specific wavelength bands are arranged. The image pickup device included in the image pickup apparatus 30 has a configuration in which pixels for photoelectrically converting light from a subject are arranged, and may be a single plate or multiple plates. The specific wavelength is a plurality of wavelength bands effective for identifying a predetermined organism (hereinafter referred to as “detection target organism”) to be detected. For example, when the organism to be detected is a predetermined cyanobacteria that produces a musty odor, the specific wavelengths are 650 nm to 700 nm (first wavelength band) and 600 nm to 650 nm (second wavelength band). The specific wavelength band may be 660 nm to 690 nm (first wavelength band) and 610 nm to 640 nm (second wavelength band). The specific wavelength band may be 670 nm to 680 nm (first wavelength band) and 620 nm to 630 nm (second wavelength band).

A pixel having sensitivity capable of discriminating signal intensity peaks in a plurality of specific wavelength bands is a plurality of wavelength bands capable of discriminating that there are signal intensity peaks in the first wavelength band and the second wavelength band. It is a pixel having sensitivity to. The plurality of wavelength bands capable of discriminating that the first wavelength band and the second wavelength band have spectral peaks are, for example, wavelengths of 550 nm to 600 nm, 600 nm to 650 nm, 650 nm to 700 nm, and 700 nm to 750 nm. It is a band.

The image pickup apparatus 30 processes signals from pixels having sensitivity in these a plurality of wavelength bands and outputs them as raw water image data. The raw water image data may be, for example, data having values of light receiving intensity in a plurality of wavelength bands in each pixel. For example, for each pixel, the value of the light receiving intensity of the light in the first wavelength band, the value of the light receiving intensity of the light in the second wavelength band, and the value of the light receiving intensity of the light in the wavelength band close to these wavelength bands. , May be given. As a result, the aquatic organism detection device 50 has spectral peaks at 650 nm to 700 nm (first wavelength band) and 600 nm to 650 nm (second wavelength band) based on the raw water image data from the image pickup device 30. It can be determined whether or not.

FIG. 2 is a diagram showing a specific configuration example of the image pickup apparatus 30. Next, with reference to FIG. 2, the configuration when the image pickup apparatus 30 is configured by using the color filter and the monochrome camera will be described. The image pickup apparatus 30 includes a color filter 31, an image pickup unit 32, and an image pickup control unit 33. The color filter 31 has a plurality of transmissive portions. Each transmitting portion allows light of a predetermined wavelength (color) to pass through and attenuates light of other wavelengths. The wavelength to be passed is different for each transmission part. For example, in the specific example of FIG. 2, the color filter 31 has six transmissive portions. The color filter 31 operates so that the transmitting portion (hereinafter referred to as “target transmitting portion”) located on the optical axis 81 of the light incident on the imaging unit 32 changes according to the control of the imaging control unit 33. In the example of FIG. 2, the color filter 31 rotates in the circumferential direction about an axis passing through the center of the disk. This rotation changes the target transmissive part. The characteristics of the transmissive portion included in the color filter 31 may be determined according to a plurality of wavelength bands effective for identifying a predetermined organism to be detected. For example, when the organism to be detected is a specific cyanobacteria, the characteristics of the transmissive portion of the color filter 31 may be determined according to the above-mentioned first wavelength band and the second wavelength band.

The image pickup unit 32 has an image pickup element. The image pickup unit 32 takes an image based on the light that reaches the image pickup element of the image pickup unit 32 centering on the optical axis 81 under the control of the image pickup control unit 33. More specifically, the image pickup unit 32 takes an image based on the light emitted by the light source 20, passing through the target transmission section of the flow cell 40, the microscope 10, and the color filter 31 and reaching the image pickup element of the image pickup section 32. The image pickup unit 32 generates data of the captured image. The image pickup unit 32 outputs the generated image data (raw water image data) to the image pickup control unit 33.

The image pickup control unit 33 controls the operations of the color filter 31 and the image pickup unit 32. For example, the imaging control unit 33 changes the target transmission unit of the color filter 31 according to a predetermined imaging timing, and causes the imaging unit 32 to perform imaging. By such control of the imaging control unit 33, an image capturing the appearance of aquatic organisms existing in the raw water passing through the flow cell 40 can be obtained based on light having a plurality of different wavelengths. For example, the image pickup control unit 33 changes the target transmission unit of the color filter 31 in a very short time, and causes the image pickup unit 32 to perform imaging in each target transmission unit. By such processing, the appearance of aquatic organisms at almost the same moment is acquired as an image of light having a plurality of different wavelengths. The image pickup control unit 33 transmits the image data acquired by the image pickup unit 32 to the aquatic organism detection device 50 via the network 60.

The explanation will be continued by returning to FIG. The flow cell 40 is an apparatus having an internal flow path through which raw water obtained from the water source to be treated passes. The flow cell 40 is composed of a transparent object so that the light of the object to be observed can reach the lens of the microscope 10. When the treatment is performed by the aquatic organism detection system 1, the raw water continues to flow in the flow path of the flow cell 40 without being almost stationary.

FIG. 3 is a diagram showing a specific example of the flow cell 40. The flow cell 40 includes a water sampling pipe 41, a main body 42, and a drainage pipe 43.

The water sampling pipe 41 has an inflow port for a part of the raw water sampled from the water source to flow into the main body 42. The raw water flowing in from the inflow port is under water pressure due to the height difference and pumps, etc., and flows in the direction of the arrow. The raw water flowing in from the inflow port of the water sampling pipe 41 passes through the water sampling pipe 41 and flows into the main body 42.

The main body 42 has a flow path inside. The main body 42 is configured by using a transparent member so that the light from the light source 20 can be uniformly transmitted regardless of the wavelength if the raw water is transparent, at least in the region to be photographed by the image pickup apparatus 30.

The drain pipe 43 has a drain port for draining the raw water that has passed through the main body 42. The raw water discharged from the drain may be disposed of in a specific area or returned to the water source.

The aquatic organism detection device 50 is configured by using an information processing device. The aquatic organism detection device 50 acquires the data of the image captured by the image pickup device 30 via the network 60. The aquatic organism detection device 50 detects the organism to be detected based on the acquired image data.

Here, a specific example of a method of selecting a specific wavelength band will be described. In the following description, specific cyanobacteria that cause musty odor are assumed as specific examples of the organisms to be detected.

Various algae including blue-green algae that generate musty odor by photographing a specific wavelength band with a hyperspectral camera capable of acquiring the signal intensity of each wavelength in 5 to 10 nm increments at a wavelength of 400 nm to 1000 nm with each pixel. Identify the spectroscopic spectrum of the algae. Then, if there is a wavelength band in which the spectral spectrum values are significantly different between the cyanobacteria that generate a musty odor and other algae, that wavelength band may be set as a specific wavelength band. Further, if there are a plurality of peak wavelengths peculiar to cyanobacteria that generate a musty odor, a plurality of wavelength bands including the plurality of peak wavelengths may be set as a specific wavelength band. The region of the pixel in which the cyanobacteria that generate a musty odor is reflected may be specified by using the value of such a specific wavelength band. In this way, the area of the pixel in which the organism to be detected is captured is called the "pixel area". By specifying the pixel region using the value of a specific wavelength band, the pixel region having a shape substantially matching the appearance shape of the organism to be detected is specified.

The blue-green algae that cause a musty odor include "Anabaena", "Oscillatoria", and "formidium", and the blue-green algae that do not cause a musty odor include "Microcystis". FIG. 4 is a diagram showing a photographic example of algae generated in raw water. In FIG. 4, Photo 61 is "Oscillatoria", which is a blue-green algae that emits a musty odor. Photo 62 is "Anabaena", a blue-green algae that produces a musty odor. Photo 63 is "formidium", which is a blue-green algae that emits a musty odor. Photo 64 shows "diatoms", which are one of the algae other than blue-green algae that do not generate a musty odor. Photo 65 is "Microcystis", a blue-green algae that does not generate a musty odor.

As shown in FIG. 4, the cyanobacteria "Anabaena", "Oscillatoria" and "Formidium" that generate a musty odor have a rod-like shape. "Microcystis", which is a blue-green algae that does not generate a musty odor, and "diatom", which is one of the algae other than the blue-green algae that does not generate a musty odor, have a spherical shape.

FIG. 5 is a diagram showing the spectral spectra of “diatom”, “microcystis”, “anabena”, “osilatoria” and “formidium”. As shown in FIG. 5, the blue-green algae "Microcystis", "Anabaena", "Oscillatoria" and "Formidium" are located near 620 nm (frame 71 in FIG. 5) and around 680 nm (frame 72 in FIG. 5). There are peaks in the spectral spectrum (downward peaks in FIG. 5). The "diatom" has spectral peaks (downward peaks in FIG. 5) at around 500 nm and around 680 nm. FIG. 5 shows a frame 71 showing a wavelength band near 620 nm and a frame 72 showing a wavelength band near 680 nm. The first wavelength band includes the wavelength band of the frame 72, and the second wavelength band includes the wavelength band of the frame 71.

It is considered that the peak around 620 nm appears mainly due to the absorption of light by a pigment called phycocyanin. It is considered that the peak near 680 nm appears mainly due to the absorption of light by the dye called chlorophyll a. In this way, the wavelength band corresponding to the dye contained in the algae becomes the peak of the spectral spectrum. Therefore, if there is a pigment peculiar to algae that generates an offensive odor such as musty odor, the wavelength band of light absorbed by the pigment becomes the wavelength band peculiar to the algae. Few algae other than blue-green algae also have phycocyanin, and phycocyanin can be said to be a pigment peculiar to blue-green algae.

Considering the above characteristics of algae, first, in the raw water image data output from the image pickup apparatus 30, the second wavelength band including the peak wavelength band of 620 nm and the first wavelength band of 680 nm, which are peculiar to blue-green algae, are included. Whether or not it is a pixel region of cyanobacteria (organisms classified including the organism to be detected) is specified by using the signal intensity in the wavelength band of. Next, image processing is performed on the raw water image data output from the image pickup apparatus 30, and the pixel region of the blue-green algae (the organism to be detected) that generates a musty odor according to the shape of the identified algae is defined as the pixel region of the blue-green algae. Can be identified. The aquatic organism detection device 50 of the present embodiment has a rod shape based on the shape of the blue-green algae pixel region after identifying the blue-green algae pixel region by the wavelength band based on the raw water image data output from the image pickup device 30. A process is performed to discriminate the pixel region of the blue-green algae from the pixel region of the blue-green algae that generates a musty odor.

In the above explanation, specific cyanobacteria that cause musty odor were assumed as the organisms to be detected. However, in the aquatic organism detection system 1 of the present embodiment, not only the above-mentioned specific blue-green algae but also any organism in water may be the detection target organism. For example, in a water treatment facility such as a water supply system, fine microorganisms (for example, diatoms) have the potential to cause damage to the filter. In addition, in water treatment facilities such as water supply systems, Microcystis has the potential to cause aggregation disorders. Therefore, in order to determine the possibility of occurrence of such a problem, aquatic organisms that can cause these problems may be set as detection target organisms.

FIG. 6 is a diagram showing a configuration example of the aquatic organism detection device 50. Next, a configuration example of the aquatic organism detection device 50 will be described. The aquatic organism detection device 50 includes an image input unit 51, a wavelength band processing unit 52, a shape discrimination processing unit 53, and a display control unit 54.

The image input unit 51 communicates with the image pickup apparatus 30 via the network 60. The image input unit 51 acquires raw water image data by communicating with the image pickup device 30.

The wavelength band processing unit 52 is a pixel of an image estimated to be a detection target organism from the image of the raw water image data based on the signal intensity of the wavelength band of each pixel included in the raw water image data acquired by the image input unit 51. Extract the area. The wavelength band processing unit 52 outputs information regarding the extracted pixel region to the shape determination processing unit 53. For example, the wavelength band processing unit 52 is given a first value (for example, “1”) to the pixel region of the image presumed to be the organism to be detected, and a second value (for example, “0”) to the other pixel regions. You may generate binary image data given ").

Hereinafter, a more specific example of processing by the wavelength band processing unit 52 will be described. The wavelength band processing unit 52 extracts a pixel region (for example, a pixel region of blue-green algae) including an image of an organism to be detected from raw water image data based on the signal intensities of the first wavelength band and the second wavelength band. The wavelength band processing unit 52 identifies, for example, pixels having peaks in the first wavelength band and the second wavelength band, and extracts the region of the specified pixel as a pixel region including an image of the organism to be detected.

The wavelength band processing unit 52 uses the first and second wavelength band processing units 52, for example, when the signal strength of the first and second wavelength bands is lower than the signal strength of the wavelength bands before and after the first and second wavelength bands. It is judged that the peak is in the wavelength band. Here, the front and rear wavelength bands are a wavelength band having a wavelength shorter than the first and second wavelength bands and a wavelength band having a wavelength longer than the first and second wavelength bands. Further, in some cases, the wavelength band processing unit 52 may perform a first wavelength when, for example, the signal strength of the first wavelength band is lower than the signal strength of the wavelength band before or after the first wavelength band. It may be determined that the band has a peak. The above-mentioned peak determination method is an example, and the wavelength band processing unit 52 determines the peaks in the first wavelength band and the second wavelength band for each pixel based on the change in the signal intensity for each wavelength band. Judge the presence or absence.

Further, the extraction process of the pixel region including the image of the organism to be detected by the wavelength band processing unit 52 is not limited to the process of extracting based on the presence or absence of peaks in the first wavelength band and the second wavelength band. For example, the extraction process shown below may be used. The wavelength band processing unit 52 calculates the detection target index using the following (Equation 1) based on, for example, the pixel data A of the first wavelength band and the pixel data B of the second wavelength band.
Detection target index = (mA-nB) / (mA + nB) ... (Equation 1)

However, m and n are arbitrary coefficients, and for the purpose of improving the accuracy of the detection target index, for example, they are coefficients that take a value from -1 to 1 (-1 ≦ m ≦ 1, -1 ≦ n ≦ 1). ). The coefficients m and n are the organisms to be detected (for example, cyanobacteria) based on the results of changing the values of the coefficients m and n by inputting the values of A and B actually measured for a plurality of types of aquatic organisms. Find the value that can distinguish between and other aquatic organisms with the highest accuracy. In this case, the wavelength band processing unit 52 determines whether or not the pixel region is the detection target based on the calculated detection target index, and extracts the detection target pixel region.

The shape determination processing unit 53 determines the shape of the pixel region of the detection target extracted by the wavelength band processing unit 52, and extracts the pixel region of the aquatic organism to be detected. For example, when the detection target is the rod-shaped cyanobacteria of the above-mentioned cyanobacteria that generate a musty odor (Oscillatoria, Anabaena, Formidium in FIG. 4), these pixel regions are extracted. Specifically, the shape discrimination processing unit 53 discriminates whether it is a rod-shaped blue-green algae shown in FIGS. 61 to 63 or a spherical blue-green algae shown in Photo 65. The discrimination method in the shape discrimination processing unit 53 can be realized by using known image processing such as pattern matching using a rod-shaped template and a spherical template. Further, the shape discrimination processing unit 53 may discriminate whether the same rod-shaped blue-green algae is oscillatoria, anabaena, or formidium by using pattern matching or a discriminator.

The display control unit 54 generates video data showing the pixel region extracted by the shape determination processing unit 53 in a mode different from that of other pixel regions. The display control unit 54 causes the display device 80 to display video data. The display control unit 54 may display not only the above-mentioned information on the pixel region but also any other information on the display device 80 as long as it is information on the detected aquatic organism.

The display device 80 is an image display device such as a liquid crystal display or an organic EL display. The display device 80 may be connected to the underwater organism detection device 50 by using a cable that transmits a video signal, or may be connected to the underwater organism detection device 50 via a network. The display device 80 displays the video data generated by the display control unit 54.

Next, the operation of the underwater organism detection device 50 will be described.
FIG. 7 is a flow chart illustrating the operation of the underwater organism detection device 50. The image input unit 51 acquires raw water image data from the image pickup device 30 (step S101). The wavelength band processing unit 52 identifies pixels having peaks in the first wavelength band and the second wavelength band based on the raw water image data input to the image input unit 51, and detects a region of the specified pixels. It is extracted as a pixel region of an image of the target organism (step S102).

The shape determination processing unit 53 determines the shape of the pixel region extracted by the wavelength band processing unit 52, and extracts the pixel region of the organism to be detected (step S103). The display control unit 54 causes the display device 80 to display information on the extracted organism to be detected (for example, an image of the extracted pixel region) (step S104).

FIG. 8 is a diagram showing a specific example of an image generated by the aquatic organism detection system 1. The image 91 is RGB data of the raw water image data captured by the image pickup apparatus 30. It should be noted that the image pickup apparatus 30 does not necessarily have to acquire RGB data. The image 92 is an image showing the extraction result by the wavelength band processing unit 52. In the image 92, the white region is a region extracted as a pixel region including an image of the organism to be detected. The image 93 is an image showing the discrimination result by the shape discrimination processing unit 53. In the image 93, the white area is the area determined as the image of the organism to be detected.

In the aquatic organism detection system 1 of the embodiment described above, the raw water image data of the raw water flowing in the flow cell 40 is photographed by the imaging device 30, and the pixel region of the image of the aquatic organism to be detected is taken from the photographed raw water data. Is shape-determined. Therefore, it is not necessary for the inspection worker to bother to collect raw water from the water source (lake, river, etc.) to be detected and prepare an inspection instrument such as a slide. Therefore, even an unskilled person can detect organisms in water in a shorter time. More specifically, it is possible to significantly reduce the time required for one inspection, and it is possible to detect aquatic organisms by targeting the raw water flowing in from the water source as an inspection target in almost real time.

Further, in the underwater organism detection system 1 of the embodiment, first, based on the information regarding the wavelength spectrum of each pixel, it is determined whether or not the region includes an image of the organism to be detected. Then, the pixel region determined to contain the image of the organism to be detected is determined based on the shape. In the shape-based discrimination process, information other than the shape is not required in the image to be processed. Therefore, the data size of the image to be processed and the processing load can be reduced. Therefore, it is possible to detect aquatic organisms in almost real time in the raw water flowing in the flow cell 40 as described above.

Further, in the underwater organism detection system 1 of the embodiment, the underwater organism detection device 50 and the display device 80 can be installed at a place away from the water source. Therefore, the degree of freedom in installing the aquatic organism detection system 1 is improved. For example, since the user can acquire the information of the detection result of the organism to be detected at a position away from the water source, it is possible to take a quick response thereafter.

(Modification example)
The wavelength band processing unit 52 may be configured to extract a pixel region including an image of an organism to be detected by using a classifier generated in advance by machine learning. In this case, the wavelength band processing unit 52 operates as follows. First, the spectral pattern of the image of the organism to be detected is acquired in advance. By using a machine learning algorithm such as SVM (Support Vector Machine), a spectrum pattern peculiar to the image of the organism to be detected is learned. In other words, a classifier that identifies the spectral pattern peculiar to the image of the organism to be detected is generated in advance. The wavelength band processing unit 52 may extract the pixel region including the image of the organism to be detected by using the discriminator generated in advance in this way. In another method as well, the wavelength band processing unit 52 may extract the pixel region by discriminating the spectrum pattern peculiar to the image of the organism to be detected based on the statistical processing.

The shape discrimination processing unit 53 may be configured to discriminate the pixel region of the organism to be detected by using a discriminator generated in advance by machine learning. In this case, the shape determination processing unit 53 operates as follows. First, an image of the organism to be detected is acquired in advance. By using a machine learning algorithm such as SVM, the shape peculiar to the image of the organism to be detected is learned. In other words, a classifier that identifies the shape peculiar to the image of the organism to be detected is generated in advance. The shape discrimination processing unit 53 may extract the pixel region of the organism to be detected by using the classifier generated in advance in this way. In another method as well, the shape discrimination processing unit 53 may extract the pixel region by discriminating the shape peculiar to the image of the organism to be detected based on the statistical processing.

Each function of the aquatic organism detection device 50 may be provided in the image pickup device 30. Further, each function of the underwater organism detection device 50 may be implemented in a portable information processing device such as a smartphone. All or part of each function of the aquatic organism detection device 50 may be implemented in a server device on the network or in the cloud.

The aquatic organism detection device 50 may be configured to store information about the detected organism to be detected in a non-volatile storage medium. Specific examples of the accumulated information include information on raw water (information on the place where water was sampled, date and time, weather, etc.), information indicating the type of organism detected, and information indicating the amount of organisms detected. ..

According to at least one embodiment described above, the aquatic organism detection device 50 detects the organism to be detected in the water flowing in the flow cell 40. At this time, the aquatic organism detection device 50 discriminates the organism to be detected based on the spectral information of the acquired image and the shape of the pixel region extracted based on the spectral information. Therefore, it is possible to detect the organism to be detected in the raw water in a shorter time.

Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, as well as in the scope of the invention described in the claims and the equivalent scope thereof.

1 ... Underwater biological detection system, 10 ... Microscope, 20 ... Light source, 30 ... Imaging device, 40 ... Flow cell, 50 ... Underwater biological detection device, 60 ... Network, 31 ... Color filter, 32 ... Imaging unit, 33 ... Imaging control unit , 41 ... Water sampling pipe, 42 ... Main body, 43 ... Drainage pipe, 51 ... Image input unit, 52 ... Wavelength band processing unit, 53 ... Shape discrimination processing unit, 54 ... Display control unit, 80 ... Display device

Claims (4)

An image input unit that acquires image data generated based on signal intensities for each of a plurality of wavelength bands for light from the raw water from an image pickup device that images a flow cell through which raw water passes.
A wavelength band processing unit that extracts a pixel region including an image of algae to be detected contained in the raw water from the image of the image data based on the transmittance of each of the plurality of wavelength bands in the image data.
A shape discrimination processing unit that discriminates the algae to be detected based on the shape of the pixel region extracted by the wavelength band processing unit.
Bei obtain aquatic organisms detection device. A flow cell through which raw water obtained from a water source flows,
A microscope that magnifies the image of the raw water flowing through the flow cell,
An imaging device that converts an image magnified by the microscope into image data based on signal intensities for each of a plurality of wavelength bands.
The aquatic organism detection device according to claim 1 and
An aquatic organism detection system equipped with. The aquatic organism detection system according to claim 2, wherein the image pickup apparatus transmits the image data to the aquatic organism detection apparatus via a network. An image input step of acquiring image data generated based on signal intensities for each of a plurality of wavelength bands of light from the raw water from an image pickup device that images a flow cell through which the raw water passes.
A wavelength band processing step of extracting a pixel region including an image of algae to be detected contained in the raw water from the image of the image data based on the transmittance of each of the plurality of wavelength bands in the image data.
A shape discrimination processing step for discriminating the algae to be detected based on the shape of the pixel region extracted in the wavelength band processing step, and a shape discrimination processing step.
Aquatic organisms detection method to have a.

Images