Analytical device, analytical system, analytical method, and program

Abstract

To efficiently support determination by a person, about a microbe culture result.SOLUTION: An analyzer comprises: means for acquiring a plurality of first images acquired by first visible light in time series, in a culture step, imaging inside of a collection container of microbes in a monitored environment; first identification means for, detecting presence of colonies of microbes or generated colonies in time series, from change of the plurality of first images in time series, then identifying shapes of the colonies whose generation or presence is detected; second identification means for identifying shapes of the colonies of the microbes, from a second image obtained by imaging inside of the collection container, by second visible light; and means for displaying states of the colonies of the microbes, on the basis of the shapes identified by the first identification means and the shapes identified by the second identification means.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an analysis device, an analysis system, an analysis method, and a program. [Background technology]

[0002] In the manufacturing environment of cell culture products or bioproducts, or in the research and development environment of these products, it may be necessary to ensure that the products are manufactured safely. In such cases, it is required to carry out predetermined environmental monitoring. In environmental monitoring, airborne and attached bacteria in the environment where the production is carried out are cultured and observed to ensure that the bacterial environment is maintained below a predetermined standard. Furthermore, environmental monitoring ultimately requires visual judgment by humans. Therefore, various image processing software-related technologies have been proposed to improve the monitoring accuracy of culture results. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 7-282220 Summary of the Invention [Problem to be solved by the invention]

[0004] However, an image that is easy for a person to judge does not necessarily match an image that is suitable for processing by a computer or other device such as image processing software. Therefore, there is room for improvement in the way in which analytical results provided by a computer or other device in environmental monitoring are presented. Therefore, an object of the present invention is to provide an analytical device, an analytical method, and a program that can effectively support a person's judgment on the microbial culture results. [Means for solving the problem]

[0005] One aspect of the present invention is exemplified by an analysis device that analyzes a collection container for collecting microorganisms. This analysis device includes a means for acquiring a plurality of first images taken in a time series with a first visible light in a culture process inside a collection container in which a microorganism collection process has been performed in a monitored environment, a first identification means for detecting the occurrence of a microbial colony or the presence of a colony that has already occurred in the time series from a change in the plurality of first images in the time series and identifying a shape of the colony whose occurrence or presence has been detected, a second identification means for identifying a shape of the microbial colony from a second image taken inside the collection container with a second visible light, and a means for displaying a state of the microbial colony based on the shape identified by the first identification means and the shape identified by the second identification means.

[0006] The first identification means detects the occurrence of a microbial colony or the presence of colonies that have already occurred over time from a change in the first image over time and identifies the shape of the colonies, where the first image is captured using a first visible light suitable for processing the first image.

[0007] The second identification means identifies the shape of the microbial colony from a second image captured with a second visible light different from the first visible light. The second image is captured with a second visible light suitable for processing the second image. Therefore, the present analysis device can display the state of the microbial colony for these two types of images based on the result of detecting colony occurrence in time series in the first image and the result of identification from the second image. Effect of the Invention

[0008] The present analytical device can effectively assist a person in making a judgment on the results of microbial culture. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram illustrating the configuration of an analysis device. [Diagram 2] FIG. 2 is a flowchart illustrating an example of the analysis process. [Diagram 3] FIG. 3 is a flowchart illustrating an example of an analysis process. [Figure 4] FIG. 4 is a flowchart illustrating an example of the analysis process. [Diagram 5] FIG. 5 is a diagram illustrating the details of the light amount adjustment process. [Figure 6] FIG. 6 is a diagram illustrating the details of the detection process for existing colonies and new colonies. [Figure 7] FIG. 7 is a diagram illustrating time-series images of the process in which colonies are generated and grow according to the culture process. [Figure 8] FIG. 8 is a diagram illustrating a process for aligning the center position of the petri dish PD. [Figure 9] FIG. 9 is a diagram illustrating a process for adjusting the rotation angle of the petri dish PD. [Figure 10] FIG. 10 is a diagram illustrating a process for classifying colonies in the current image into existing colonies and new colonies. [Figure 11] FIG. 11 is a diagram illustrating an environment determination system according to the second embodiment. [Figure 12] FIG. 12 is a diagram illustrating an environment monitoring system according to the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] An analysis device and an analysis method according to an embodiment will be described below with reference to the drawings.

[0011] <First embodiment> An analysis device 10 according to a first embodiment will be described with reference to Figs. 1 to 10. The analysis device 10 analyzes a petri dish PD that holds a sample acquired in a monitored environment, such as a manufacturing environment for a cell culture product or a bioproduct, or a research and development environment. The monitored environment is also called a sterile room or a sterilization room. The petri dish PD is also called a collection container. The petri dish PD is a transparent, disk-shaped container also called a Petri dish. The petri dish PD has a shallow cylindrical or dish-shaped main body with a transparent bottom and an open top, and a transparent lid that covers and closes the open part of the main body.

[0012] The petri dish PD is placed in a predetermined position in the monitored environment with the top open except for the lid, and obtains samples to be inspected of microorganisms floating or falling in the air, such as those called suspended bacteria (also called falling bacteria). The petri dish PD also obtains samples to be inspected of microorganisms attached to objects placed in the monitored environment or to the human body or clothing of workers working in the monitored environment, such as those called attached bacteria. For microorganisms attached to the human body or clothing, the sample to be inspected is collected by contacting the bottom surface of the inner wall of the petri dish PD with the location to be inspected.

[0013] The petri dish PD from which the test sample has been obtained is placed in a culture process in which microorganisms such as bacteria and mold are cultured. The analysis device 10 obtains and analyzes images from the petri dish PD in the culture process at predetermined intervals. The images are, for example, images of the inside bottom surface of the petri dish PD taken through a transparent lid. If the obtained test sample contains more than a predetermined amount of microorganisms, the microorganisms will gradually develop according to the culture process. The microorganisms that have developed into a mass visible to the human eye are called colonies.

[0014] The analysis device 10 of this embodiment detects colonies of microorganisms from images of the petri dish PD during or after the completion of the culture in the culture process, and monitors the environment by a person (also called an operator or user). Support for trolling.

[0015] (composition) FIG. 1 is a diagram illustrating the configuration of an analysis device 10. As shown in FIG. 1, the analysis device 10 mainly has an information processing unit, an information acquisition unit, and an imaging unit. Although omitted in FIG. 1, the petri dish PD is placed at a position where the bottom surface of the inner wall of the petri dish PD can be photographed by the imaging unit through a transparent lid. In this case, the analysis device 10 may have a mounting table for placing the petri dish PD. The petri dish PD may also be held at a position where the bottom surface of the inner wall of the petri dish PD can be photographed by the imaging unit through a transparent lid. In this case, the analysis device 10 may have a holder for holding the petri dish PD.

[0016] The information processing section of the analysis device 10 has a central processing unit (CPU 11), a main memory section 12, and an external device connected through an interface (I / F), and executes information processing by a program. The CPU 11 is also called a processor. Examples of the external device include an external memory section 13, a display section 14, an operation section 15, and a communication section 16. The information processing section may be a computer incorporated in the housing of the analysis device 10, or may be a normal personal computer or the like that is separate from the housing of the analysis device 10.

[0017] The CPU 11 executes a computer program that has been loaded in an executable manner in the main memory unit 12, and provides the functions of the analysis device 10. The CPU 11 is also called a processor. The CPU 11 is not limited to a single processor, and may have a multi-processor configuration. The CPU 11 may also include a graphics processing unit (GPU), a digital signal processor (DSP), etc. The CPU 11 may also include a field programmable gate array ( The main memory unit 12 stores the computer program executed by the CPU 11, the data processed by the CPU 11, and the like.

[0018] The main memory unit 12 is a Dynamic Random Access Memory (DRAM), a Static Random Access The external storage unit 13 may be a static random access memory (SRAM), a read only memory (ROM), etc. is used, for example, as a storage area supporting the main storage unit 12, and stores computer programs executed by the CPU 11, data processed by the CPU 11, etc. The external storage unit 13 is a hard disk drive, a solid state disk (SSD), etc. Furthermore, the analysis device 10 may be provided with a drive device for a removable storage medium. The removable storage medium is, for example, a Blu-ray disc, a Digital Versatile Disk (DVD), a Compact Disc (CD), a flash memory card, a cloud, etc. Here, the cloud is an external storage device connected via a network.

[0019] The display unit 14 is, for example, a liquid crystal display, an electroluminescence panel, etc. The operation unit 15 is, for example, a keyboard, a pointing device, etc. In this embodiment, examples of the pointing device include a mouse, a touch pad, and a touch panel. The communication unit 16 exchanges data with other devices on the network.

[0020] The information acquisition unit of the analysis device 10 is connected to the information processing unit via a predetermined interface. The information acquisition unit controls the photographing unit, adjusts the photographing environment, acquires images from the photographing unit, and transfers them to the information processing unit. The information acquisition unit has a camera control unit 17, a white light control unit 18, a blue light control unit 19, an LED control unit 20, and a fan control unit 21. The photographing unit projects illumination light onto the petri dish PD, removes dust adhering to the surface of the petri dish PD, sets photographing conditions, and photographs an image of the petri dish PD. The photographing unit has a camera CM, a white light LW, a blue light LB, LEDs L1 to L8, and a fan FN.

[0021] The camera control unit 17 acquires an image from the camera CM and stores the acquired image in an image memory. Then, the camera control unit 17 transfers the image in the image memory to the CPU 11. The camera CM is a Charge Coupled Device (CCD) camera, a Complementary Metal Oxide Semiconductor (CMOS) camera, or the like. The camera CM transfers image data to the camera control unit 17 at a predetermined frame period, and causes the image data to be stored in the image memory.

[0022] The white light control unit 18 controls the on / off of the white light LW and the amount of white light emitted from the white light LW according to a command from the CPU 11. The white light LW, for example, irradiates a phosphor with light from a blue light emitting diode (LED) to emit white light. The white light LW has a semiconductor switch, which is turned on by a control signal from the white light control unit 18 to emit white light. The white light LW also has, for example, a dimming circuit. The dimming circuit has, for example, a rectifier circuit and a DC (direct current) to DC converter, and controls the amount of current flowing through the blue LED by controlling the pulse width (duty ratio) of the DCDC converter to adjust the amount of light emitted from the white light LW. The white light control unit 18 outputs a signal to command the semiconductor switch to be on or off and a signal to specify the pulse width (duty ratio) of the DCDC converter to the white light LW. The white light control unit 18 may have a dimming circuit and supply power to the white light LW while controlling the current value to the white light LW.

[0023] The blue light control unit 19 controls the on / off of the blue light LB and the amount of blue light irradiated from the blue light LB in response to a command from the CPU 11. The blue light LB includes, for example, a blue LED or an array of blue LEDs. The blue light LB may have a phosphor that emits light of a blue wavelength. The blue light LB irradiates the blue light from below the petri dish PD. The camera CM photographs the bottom surface of the petri dish PD, onto which blue illumination light is projected from below by the blue light LB, through the transparent lid on the upper side. That is, the blue light LB irradiates the petri dish PD with blue transmitted light.

[0024] The blue light LB has a semiconductor switch and a dimming circuit similar to those of the white light LW. The blue light control unit 19 outputs a signal to the blue light LB to command the semiconductor to be on or off and a signal to specify the pulse width (duty ratio) of the DCDC converter. The blue light control unit 19 may have a dimming circuit and supply power while controlling the current value to the blue light LB. The camera CM captures a first image of the petri dish PD by the transmitted light of the blue light LB and a second image by the reflected light of the white light LW. Therefore, the camera CM is an example of a photographing means for capturing the first and second images. The image by the transmitted light from the blue light LB is an image suitable for processing by a computer program such as image processing software. This is because a single color is suitable for image processing software. The inventors of the present analysis device 10 have empirically recognized that the accuracy of colony detection is higher in images by the blue light LB than in the case of other lights.

[0025] The LED control unit 20 controls the eight LEDs L1 to L8 to be on and off. Although omitted in FIG. 1, the LEDs L1 to L8 are arranged above the petri dish PD so as to illuminate the vicinity of the center of the petri dish PD at an angle of 45 degrees in a circular shape. The LEDs L1 to L8 each emit directional white light in a predetermined direction. Therefore, the LEDs L1 to L8 irradiate directional light from eight directions on the upper side of the petri dish PD through the lid toward the center of the inner wall bottom surface. Note that the number of LEDs L1, etc. is not limited to eight. For example, the photographing unit may be provided with four LEDs from four directions. The eight LEDs L1 to L8 act to photograph the colonies of microorganisms in the petri dish PD in a three-dimensional manner by directional white light from different directions. In this way, the analysis device 10 assists the user in determining whether the shadow in the image of the petri dish PD is a colony with a three-dimensional structure or a flat stain, cloudiness, etc. The white light of LEDs L1 to L8 is emitted by a blue LED and a phosphor, as is the white light LW.

[0026] The fan control unit 21 controls the on / off of the fan FN. When the fan FN is turned on, it rotates its blades and blows a predetermined amount of air (thick dotted arrow in FIG. 1) onto the petri dish PD. The analysis device 10 uses the action of the fan FN to remove foreign matter such as dust attached to the surface of the petri dish PD placed or held under the camera CM. Before photographing with the camera CM, the CPU 11 instructs the fan control unit 21 to turn on the fan FN or to repeat on and off a predetermined number of times to remove foreign matter on the surface of the petri dish PD. It is preferable that the fan FN be capable of blowing air onto the surface of the petri dish PD, particularly onto both the top surface of the lid and the bottom surface of the main body. The fan FN is an example of a blowing means. That is, the fan FN blows air onto the surface of a sampling container that is placed in a state where it can be photographed by the camera CM, which is the photographing means.

[0027] Although not illustrated in FIG. 1, a transport mechanism may be provided that transports the petri dish PD placed in the culture process to a mounting table or holder at a position where the photographing unit of the analysis device 10 can photograph the dish.

[0028] (process) 2 to 6, the processing procedure of the analysis method in the analysis device 10 will be illustrated. These processes are executed by the CPU 11 using a computer program executablely deployed in the main storage unit 12. However, in the following description, these processes will be mainly described as being executed by the analysis device 10. Among these, FIGS. 2 to 4 are flow charts illustrating the processing procedure of the analysis. The processes illustrated in FIGS. 2 to 4 are executed a predetermined number of times at predetermined timings in the culture process, with the first process being before the petri dish PD enters the culture process.

[0029] In this process, first, the analysis device 10 transports the petri dish PD placed in the culture process, and sets it in a state where the bottom surface of the inner wall of the petri dish PD can be photographed by the imaging unit (S21). However, if the analysis device 10 does not have a transport mechanism, the petri dish PD is set in a state where the bottom surface of the inner wall of the petri dish PD can be photographed by the intervention of the operator. In this case, the analysis device 10 recognizes that the setting of the petri dish PD has been completed and that it is in a state where it can be analyzed, by the operator's input through the operation unit 15. Alternatively, the analysis device 10 may recognize that the setting of the petri dish PD has been completed and that it is in a state where it can be analyzed, by a signal from a switch provided on the mounting base on which the petri dish PD is placed, a switch provided on the holder that holds the petri dish PD, or the like.

[0030] Next, the analysis device 10 operates the fan FN by the fan control unit 21 to blow air toward the petri dish PD and remove foreign matter on the surface of the petri dish PD (S22). Next, the analysis device 10 adjusts the light intensity of the white light LW and the blue light LB (S23). The process of adjusting the light intensity will be described in detail with reference to FIG. 5.

[0031] Next, the analysis device 10 photographs the petri dish PD with the camera CM while irradiating the petri dish PD with transmitted light from below by blue light LB (S24). The amount of transmitted light is adjusted in S23. The analysis device 10 repeats the processes of Figs. 2 to 4 multiple times in chronological order as the culture process progresses. Therefore, the process of S24 is an example of acquiring multiple first images of the inside of the petri dish PD, which is a collection container, in chronological order during the culture process using the first visible light, which is blue light. The CPU 11 of the analysis device 10 acquires the images acquired in the process of S24 multiple times in chronological order via the camera control unit 17. Therefore, the CPU 11 can also be said to be a means for acquiring multiple first images.

[0032] Next, the analysis device 10 captures an image of the petri dish PD with the camera CM while irradiating the petri dish PD with light from above using the white light LW (S25). An image is captured using the reflected light of the white light emitted from the white light source LW. The amount of white light emitted from the white light source LW is adjusted by S23.

[0033] Next, the analysis device 10 performs the following processes of S26 to S42 on the image captured under irradiation with transmitted light by the blue light LB. That is, the analysis device 10 performs alignment between the image captured in the previous culture step (previous image) and the image captured this time with blue light (present image) by the process of S24 (S26). The previous image also includes an image before culture. For example, the main memory unit 12 stores an image before culture captured in the first process before the petri dish PD enters the culture step, and images captured in the subsequent processes up to the previous time. In the processes of Figs. 2 to 4, a process of detecting the occurrence of a colony or a process of identifying a detected colony is performed by comparing the captured image with the present image. As a pre-processing before performing such a comparison between images, the analysis device 10 performs alignment between the images.

[0034] 8 and 9, marks M1 and M2 are formed on the edge of the upper surface of the petri dish PD. The alignment process includes a process of aligning the centers of the petri dishes and a process of rotating one of the two images so that the marks M1 and M2 on the edge overlap.

[0035] Next, the analysis device 10 executes a process of detecting colonies in the aligned previous image and the current image (S27). This process is, for example, a process of detecting a shadow having a predetermined dimension or more in the current image. The predetermined dimension is, for example, stored in the main memory unit 12 as a system parameter. Note that since the colonies in the previous image have already been detected in the previous process, the analysis device 10 need only read out the detection result in the previous image from the main memory unit 12.

[0036] Next, the analysis device 10 removes the initial foreign matter from the current image (S28). The analysis device 10 stores in the main memory unit 12 the pre-culture image captured in the first process before the petri dish PD enters the culture process. The shadow in the pre-culture image is the shadow of something other than a colony, such as a foreign matter or noise. Therefore, if the current image contains a shadow identical to the shadow in the pre-culture image, the analysis device 10 removes the shadow (S28). The removal of the identical shadow can be performed by subtracting the shadow in the pre-culture image from the shadow in the current image. The process of removing the identical shadow is provided by image processing software mounted on the CPU 11. Then, the analysis device 10 stores in the main memory unit 12 the current image from which the initial foreign matter has been removed (the current colony detection image in FIG. 2). Note that the previous image from which the initial foreign matter has been removed (the previous colony detection image in FIG. 2) has already been stored in the main memory unit 12.

[0037] Next, the process will be described with reference to FIG. 3. The arrow A in FIG. 2 indicates that the process continues to the arrow A in FIG. 3. The analysis device 10 compares two images from which initial foreign matter has been removed, that is, the previous image and the current image (S31). Then, if the shadow that was in the previous image has disappeared in the current image (Y in S32), the analysis device 10 determines that the disappeared shadow is a foreign matter and deletes the colony (polygon) at the corresponding location in the detection result based on the previous image (S33). The analysis device 10 also subtracts the number of colonies corresponding to the disappeared shadow from the number of colonies detected up to the previous time. The process of S33 is an example of a process of excluding the shadow that was in the previous image as a shadow other than a microorganism colony when the shadow that was in the previous image does not exist in the subsequent image among the images before and after in the time series. The judgment in S32 is an example of a process of distinguishing between the shadow of a microorganism colony and the shadow other than a microorganism colony. Therefore, the CPU 11 of the analysis device 10 executes the judgment in S32 as an example of a discrimination means.

[0038] However, if it is difficult to compare the previous image with the current image (Y in S34), the analysis device 10 outputs a message to the display unit 14 and prompts the operator to input the discrimination result (S35). In this case, the analysis device 10 may output a stereoscopic image in which light is irradiated from eight directions by LEDs L1 to L8 to the display unit 14. If it is difficult to make a judgment in S34, it can be said that it is difficult to make a judgment by the discrimination means. The stereoscopic image in which light is irradiated from eight directions is an example of an image captured in a state in which the petri dish PD, which is a collection container, is illuminated with illumination light from multiple directions. Note that the stereoscopic image may be an image in which light is irradiated from directions other than eight, for example, from four directions.

[0039] Here, the case where it is difficult to judge is when a new shadow appears in the image of the last colony detection process after the culture process is completed, and it is not possible to confirm whether the newly appeared shadow will disappear next time. The analysis device 10 holds, for example, information on whether this is the last colony detection process or information on the number of times the colony detection process is executed in the culture process. Therefore, the analysis device 10 can recognize whether the current image is the last colony detection process image after the culture process is completed. In addition, an example of a case where it is difficult to judge is when, for example, there is a shadow in the current image that partially overlaps with a shadow in the previous image, making it difficult to judge whether the shadow has disappeared. In such a case, the judgment of whether the shadow is a foreign object or not is decided by the input of the discrimination result by the operator. At this time, the analysis device 10 supports the operator's judgment by a stereoscopic image irradiated with light from eight directions by the LEDs L1 to L8. The process of S35 is an example of a process of presenting information that prompts the operator to intervene. The CPU 11 of the analysis device 10 executes the process of S35 as an example of a presentation means.

[0040] Next, the analysis device 10 executes a process for detecting existing colonies and new colonies (S36). The process for detecting existing colonies and new colonies will be described separately with reference to FIG. 6. Here, the process is executed on images (images taken under irradiation with transmitted light by blue light LB) taken at each stage of the culture process and stored in chronological order. That is, the analysis device 10 compares each of the images stored in chronological order with the previous image to identify newly generated colonies and existing colonies, and extracts the contours and centers of gravity of each.

[0041] Next, the processing will be described with reference to Fig. 4. Arrow B in Fig. 3 indicates that the processing continues to arrow B in Fig. 4. The analysis device 10 determines whether or not a new colony has occurred in the current image as a result of the processing of S36 (S41). If a new colony has occurred in the current image, the number of new colonies is added to the number of colonies.

[0042] Next, the analysis device 10 extracts a contour from an image captured with reflected light of white light irradiated from the white light LW (S43). The process of S43 can be said to be a process of identifying the shape of a microbial colony from a second image captured with white light, which is the second visible light, inside the petri dish PD, which is a collection container. This process is also performed at each stage of the culture process. Therefore, the process of S43 is an example of a process of identifying the shape of a microbial colony from a second image captured with white light, which is the second visible light, inside the petri dish PD, which is a collection container. In this embodiment, the CPU 11 of the analysis device 10 that executes the detection process of existing colonies and new colonies in FIG. 6 is called the first identification means, and the process of FIG. 6 is called the first step. This is a process of detecting colonies from an image captured with blue transmitted light. On the other hand, the CPU 11 of the analysis device 10 that executes the process of S43 is called the second identification means, and the process of S43 is called the second step.

[0043] Next, the analysis device 10 executes a process of matching the colony contours acquired in the detection process of existing colonies and new colonies in S36, and displays the processing results on the display unit 14 (S44). Here, the analysis device 10 determines the colony contour from the contours extracted from previous and next images acquired in time series using blue transmitted light, and the contour extracted from the image captured using reflected white light in the process of S43. For example, the analysis device 10 displays the colony contour acquired in the detection process of S36 and the contour extracted in the process of S43 superimposed on each other. However, if there is a mismatch between the contour in the detection process of S36 and the contour in the process of S43, the analysis device 10 may give priority to the contour in the detection process of S36 for the mismatched parts. This is because the detection process of S36 This is because in the extraction process, the contour is acquired from a series of images taken at each stage of the culture process and stored in chronological order, so the reliability of the detection result is high. Also, in the detection process of S36, images taken under irradiation with transmitted light by blue light LB are used, so that the reliability of the image processing result is empirically high. For example, the analysis device 10 may not use as a contour a portion of the contour extracted in the process of S43 that is farther away than a predetermined limit from the contour obtained by the detection process of S36.

[0044] In addition, in the image taken with reflected white light, a single outline appears, but in reality, multiple colonies may overlap. Therefore, the analysis device 10 displays the image taken with reflected white light with the outline of the colony added thereto, and holds the counted number of colonies counted in the process of S42. Therefore, the analysis device 10 displays the colonies (polygons) obtained by the time series processing of the blue light image and the colonies confirmed to match the edges of the shadows on the white light image superimposed on the image taken with reflected white light seen by the operator. The analysis device 10 also presents the number of colonies obtained by the time series processing up to that point to the operator. Here, the displayed information such as the outline of the colony, the image taken with white light, and the number of colonies can be said to be information indicating the state of the microbial colony based on the shape specified by the first specifying means (FIG. 6) and the shape specified by the second specifying means (S43). The CPU 11 of the analysis device 10 also executes the process of S44 as a means for displaying the state of the microbial colony.

[0045] Furthermore, when the operator requests a chronological display by operating the operation unit 15 of the analysis device 10 (Y in S45), the analysis device 10 executes the chronological display. In the chronological display, the analysis device 10 chronologically displays the outlines of colonies at each stage that have already been extracted at each stage of the culture process or the images that have been processed at each stage (S46). At this time, the operator can freely enlarge or reduce the outlines of colonies at each stage or the images that have been processed at each stage by operating the operation unit 15, and display them on the display unit 14. The process of S46 is an example of a process for displaying the state of colonies in chronological order. The CPU 11 of the analysis device 10 executes the process of S46 as an example of a means for displaying the state of colonies in chronological order.

[0046] Then, based on the information such as the outlines and images of the colonies and the number of counted colonies displayed in the processes of S44 to S46, the operator (also called an inspector) judges the presence or absence of colonies (S47). That is, if the number of detected colonies is less than the reference value (n), for example, 0, the operator inputs a negative judgment (S48). On the other hand, if the number of detected colonies is equal to or greater than the reference value (n), the operator inputs a positive judgment (S49). The processes of S48 and S49 can be said to be processes for accepting user confirmation regarding the presence or absence of microorganisms from the state of the displayed colonies. The CPU 11 of the analysis device 10 executes the processes of S48 and S49 as a means for accepting user confirmation. Then, the analysis device 10 overwrites and saves the number of detected colonies and the like together with the result of the positive judgment or negative judgment (S4A).

[0047] Then, the analysis device 10 judges whether the current process is the final stage process after the end of the culture process (S4B). If the current process is the final stage process, the analysis device 10 ends the process for this petri dish PD. On the other hand, if the current process is not the final stage process, the analysis device 10 returns this petri dish PD to the next stage of the culture process (S21). Then, after the next stage is completed, the analysis device 10 executes the processes from S22 onwards again. The arrow C in FIG. 4 indicates that the process continues to the arrow C in FIG. 2.

[0048] FIG. 5 shows an example of the light intensity adjustment process (S23 in FIG. 2) in detail. The light intensity is adjusted because the appropriate light intensity may differ depending on the culture state of the microorganisms in the petri dish PD. For example, if the area of ​​the bottom surface of the inner wall of the petri dish PD that is occupied by colonies is small, if the light intensity is high, the image captured by the camera CM of the imaging unit will be too bright. In this case, even if a colony is present in the petri dish PD, the shadow of the colony will not remain in the captured image. There may be cases where this is not possible.

[0049] On the other hand, for example, when the colonies occupy a large portion of the bottom surface of the inner wall of the petri dish PD, if the amount of light is small, the image captured by the camera CM of the photographing unit will be too dark. In this case, even if many colonies are present in the petri dish PD, it may be the case that the shadows of the colonies cannot be clearly identified in the captured image. Therefore, the analysis device 10 counts the number of colonies from each of multiple images captured with different amounts of light, and searches for a stable state in which the number of colonies changes little. The amount of light in this stable state is set as the desired amount of light. The following processing is performed with both the blue light LB and the white light LW. Here, the processing for the blue light LB is illustrated, but the processing is similar for the white light LW. That is, the analysis device 10 detects the number of colonies from an image in which the amount of light is gradually changed using one of the amount of blue light LB, which is the first visible light, and the amount of white light LW, which is the second visible light. The following description will be given using the blue light LB as an example.

[0050] In this process, the analysis device 10 captures images of the petri dish PD by changing the amount of blue light LB in multiple steps (S51). Here, the captured images are an example of multiple images captured by changing the amount of light in stages. Then, the analysis device 10 counts the number of patterns in each image (S52). The number of patterns is counted, for example, by converting the shadows in the image into polygons (IDs and vertex strings) and counting the number of polygons (number of IDs).

[0051] Then, the analysis device 10 confirms that there is a pattern in any of the images, that is, the number of patterns is not 0 (S53). In other words, if there are no colonies at all on the petri dish PD, the number of patterns will not change even if the light intensity is changed in multiple stages.

[0052] If the number of patterns is not 0 (Y in S53), the analysis device 10 detects the amount of light when the change in the number of patterns is equal to or less than a predetermined threshold, for example, when there is no change in the number of patterns, and determines it as the appropriate amount of light (S54). Here, the state where there is no change in the number of patterns is an example of a stage where the change in the number of shadows between multiple images captured by gradually changing the amount of light satisfies a predetermined condition. Furthermore, the analysis device 10 sets the amount of blue light LB to the appropriate amount of light determined in the process of S54 (S55). The process of S55 is an example of a process of setting the amount of light when capturing an image.

[0053] The blue light LB is an example of the first visible light, and the white light LW is an example of the second visible light. Therefore, the CPU 11 of the analysis device 10 executes the process of FIG. 5 as an example of a means for setting the light amount.

[0054] FIG. 6 is a diagram illustrating the details of the detection process for existing colonies and new colonies (S36 in FIG. 3). In this process, the analysis device 10 extracts a polygon from the contours of the current image, which comes before and after in a time series (S61). Then, the analysis device 10 calculates the center of gravity of the polygon in the current image (S62). Then, the analysis device 10 determines whether or not a colony has already been detected in the previous image (S63). If a colony has already been detected in the previous image (Y in S63), the analysis device 10 detects the occurrence of a new colony by matching the colonies already detected in the previous image with the colonies detected in the current image (S64).

[0055] That is, the analysis device 10 classifies the polygon in the current image that is closest to the contour (polygon) of each colony already extracted in the previous image relative to the current image in the time series as an existing colony. The following methods can be exemplified for classifying existing colonies.

[0056] (1) Select as an existing colony the polygon in the current image whose edge is closest to the edge of the polygon in the previous image.

[0057] (2) The polygon in the current image that has the edge closest to the center of gravity of the polygon in the previous image is selected as an existing colony.

[0058] (3) The polygon in the current image whose center of gravity is closest to the center of gravity of the polygon in the previous image is selected as an existing colony.

[0059] (4) In each of (1) to (3) above, points are assigned in order of closeness, and the polygon in the current image with the highest total points is selected.

[0060] (5) In the above (1) to (4), if a polygon in the current image contains a polygon in the previous image, the containing polygon in the current image is preferentially selected as an existing colony. The process of S64 can be said to be a process of detecting the presence of colonies that have already emerged in time series. This process can be said to detect the presence of colonies that have already emerged based on at least one of the boundary line of the shadows and the center of gravity of the area surrounded by the boundary line that exists in each of the multiple first images captured with blue light.

[0061] Furthermore, the analysis device 10 classifies colonies other than existing colonies as new colonies. On the other hand, if the occurrence of a colony has not yet been detected up to the previous image (N in S63), the analysis device 10 detects all polygons detected in the current image as new colonies (S65). The process of S65 can be said to be a process of detecting the occurrence of a microbial colony. Therefore, the process of FIG. 6 is an example of a process of identifying the shape of a colony whose occurrence or existence has been detected. In this embodiment, the CPU 11l of the analysis device 10 that executes the process of FIG. 6 is called a first identification means. Also, the process of FIG. 6 is called a first step. Also, the process of FIG. 6 is an example of a process of detecting the occurrence of a microbial colony or the existence of a colony that has already occurred based on at least one of the boundary line of the shadow present in each of the multiple first images captured with blue light and the center of gravity of the area surrounded by the boundary line.

[0062] (Processing example) 7 is a diagram illustrating time-series images of the process in which colonies are generated and grow according to the culture process. Illumination is by blue light LB transmitted from below the petri dish PD.

[0063] In Fig. 7, the arrows pointing from top to bottom indicate the passage of time. The three images in the left column of Fig. 7 are examples of raw images of the petri dish PD. The three images in the right column of Fig. 7 are examples of images in which image processing has been performed on the raw images to remove noise and the like.

[0064] The analysis device 10 first photographs the petri dish PD before the culture process begins (image line G1). In this state, shadows are usually not detected. Also, shadows detected in this state are abnormal. Therefore, the shadow of the foreign object is removed from the subsequent images. This is as described in the process of S28 in FIG. 2.

[0065] As the culture process progresses, a new colony appears (row G2 of the image). A shadow that is not present in the previous image is determined to be a new colony. That is, the analysis device 10 compares images of the petri dish PD at previous and subsequent stages in the time series of the culture process, and if there is no matching colony position information, it is determined to be a new colony. This is as described in the process of S65 in FIG. 6.

[0066] In addition, the analysis device 10 compares the images of the petri dish PD before and after the process, and if the shadow disappears in the later image (current image), it excludes it as a foreign object (dust) at that timing. This is as described in the process of S33 in FIG. 3.

[0067] Furthermore, as the culture process progresses, new colonies are additionally generated (row G3 in the image). The analysis device 10 classifies the colony in the current image that is closest to the colony in the previous image as an existing colony (see the correspondence indicated by the dotted arrow). The analysis device 10 also classifies the colony in the previous image that is not classified as an existing colony as a new colony. In row G3, existing colonies are illustrated with a hatched pattern, and new colonies are illustrated with a solid black pattern.

[0068] 8 and 9 illustrate an example of a process for aligning a previous image with a current image. FIG. 8 illustrates a process for aligning the center position of a petri dish PD. In FIG. 8, the previous image is labeled "before," and the current image is labeled "after." As shown in FIGS. 8 and 9, marks M1 and M2 are attached to the edges of the petri dish PD.

[0069] First, the analysis device 10 aligns the center position of the petri dish PD in the previous image with the center position of the petri dish PD in the current image. That is, the analysis device 10 calculates the difference between the center coordinates of the petri dish in the previous image and the current image. Then, the analysis device 10 relatively moves one of the images by the amount of the difference to align the position of the petri dish PD.

[0070] Fig. 9 illustrates a process for adjusting the rotation angle of the petri dish PD. After aligning the center position in Fig. 8, the analysis device 10 rotates the image to adjust the rotation angle. Marks M1 and M2 are attached to the edge of the petri dish PD. Therefore, the analysis device 10 calculates the reference positions of the marks M1 and M2 in the previous image and the current image, for example, the angles of the center positions of the marks M1 and M2, and rotates one of them.

[0071] Figure 10 illustrates the process of classifying colonies in the current image into existing colonies and new colonies. The image labeled "before" in Figure 10 is a polygon showing a colony obtained from the previous image in the time series. The image labeled "after" is a polygon showing a colony obtained later in the time series, i.e., from the current image. "Before + After" illustrates the state where both polygons are superimposed.

[0072] In this process, the analysis device 10 calculates the shortest distance from the previous colony (polygon) to each of the current colonies (polygons), and determines that the colony in the current image that is the closest distance to the previous colony is a grown existing colony. In the figure, the previous colony is completely contained within one of the current colonies. Also, the colony with the closest edge to the previous colony is the current colony that contains the previous colony. Therefore, the analysis device 10 determines that the current colony that contains the previous colony is an existing colony. The analysis device 10 performs the above process on all previous colonies. Then, the analysis device 10 determines that the colony in the current image that cannot be associated with any of the previous colonies is a new colony.

[0073] The analysis device 10 registers the boundary (polygon) and center of gravity of a new colony, and registers the boundary (polygon) and center of gravity of an existing colony by replacing them with those of the previous colony.

[0074] (Effects of the embodiment) In the above embodiment, the analysis device 10 compares the contours extracted between images captured in time series using blue transmitted light with the contours extracted from images captured using reflected white light. The analysis device 10 then determines that the contours of the two match as the contours of the colony. The contours extracted in time series are data acquired by repeating image processing by the CPU 11. The contours extracted in time series are also contours from images captured under irradiation with transmitted light using blue light LB, which is suitable for the image processing of the CPU 11. These contours are also extracted by comparing each image in the time series with the previous image, identifying newly developed colonies and existing colonies. On the other hand, the contours extracted from images captured using reflected white light are not easily distinguished by the human eye. The contours are obtained from an image suitable for recognition by the human eye. Therefore, the analysis device 10 compares and collates the contours of the processing result in the image that is easily visible to the human eye with the contours extracted in time series, thereby enabling accurate colony analysis and presenting the colony contours to the operator. In this case, the analysis device 10 analyzes the time series image in one process, so that even if colonies overlap during the culture process, it is possible to accurately distinguish the colonies and count the number of colonies.

[0075] In addition, in the above-mentioned time-series processing, the analysis device 10 classifies the polygon in the current image that is closest to the contour (or center of gravity) of each polygon in the previous image as an existing colony, and classifies other than existing colonies as new colonies. By comparing polygons obtained from such time-series images, the analysis device 10 can accurately distinguish existing colonies and determine the occurrence of new colonies.

[0076] When a shadow that existed in the previous image in the time series has disappeared in the current image, the analysis device 10 judges the image that has disappeared to be a foreign object, such as dust, and removes it from the previous image. Therefore, the analysis device 10 can appropriately remove the foreign object and perform the analysis. Furthermore, the analysis device 10 can count the number of colonies with high accuracy by this process.

[0077] When it is difficult to compare the current image with the previous image in the time series, the analysis device 10 outputs a message to the display unit 14 to prompt the operator to input. Therefore, the analysis device 10 can obtain the operator's decision at an appropriate time. In this case, the analysis device 10 can output a stereoscopic image in which light is irradiated from eight directions by the LEDs L1 to L8 to the display unit 14. Therefore, the analysis device 10 can display the colony three-dimensionally and assist the operator in deciding whether it is a colony or a cloudy Petri dish PD.

[0078] The analysis device 10 operates the fan FN by the fan control unit 21 to blow air toward the petri dish PD and remove foreign matter from the surface of the petri dish PD. Therefore, the analysis device 10 can perform analysis while minimizing misidentification caused by foreign matter.

[0079] The analysis device 10 receives input from an operator based on the outlines and images of colonies, and stores the operator's judgment results on the presence or absence of colonies. In other words, the analysis device 10 functions solely as a device that supports the operator's judgment.

[0080] When capturing images with either blue light LB or white light LW in multiple steps, the analysis device 10 captures images of the petri dish PD by changing the light intensity. The analysis device 10 then detects the light intensity in a stable state where the change in the number of patterns counted in each image is equal to or less than a predetermined threshold, and determines this as the appropriate light intensity. Therefore, the analysis device 10 can capture images with the appropriate light intensity for each of the blue light LB and white light LW.

[0081] The analysis device 10 displays, in chronological order, the outlines of colonies or images during processing that have already been extracted at each stage of the culture process. At this time, the operator can freely enlarge or reduce the outlines of colonies or images during processing at each stage and display them on the display unit 14 by operating the operation unit 15. Therefore, the analysis device 10 can appropriately support the operator's judgment.

[0082] (Modification) In the above embodiment, the processes in Figs. 2 to 4 are exemplified as a series of processes. However, the processes of the analysis device 10 are not limited to the above processes. For example, the analysis device 10 may execute S21 to S25 in Fig. 2, and the processes from S26 onwards may be executed by another computer connected via a network. For example, the analysis device 10 may transmit image data and the like obtained in the processes from S21 to S25 to another computer, server, cloud, or the like connected via a network. The image data may be transferred to a recording device or the like on the network. Then, another computer connected to the network may refer to the image data stored in the cloud or the like and execute the processes from S26 onwards. Even when the analysis device 10 executes the series of processes shown in Figs. 2 to 4, the image data or the like may be stored in another computer, server, or a recording device on the cloud on the network. Then, the analysis device 10 may execute the processes from S26 onwards from the data acquired in the current process and data acquired in the past and stored in another computer or the like. In that case, the analysis device 10 may download the data acquired in the past from the other computer or the like when processing the data acquired in the current process.

[0083] <Second embodiment> The second embodiment will be described with reference to Fig. 11. Fig. 11 is a diagram illustrating an environment determination system 100 of the second embodiment. The environment determination system 100 includes an analysis device 10, a DNA analysis device 40, and a determination device 50. The analysis device 10 is similar to that described in the first embodiment. That is, the analysis device 10 inputs information such as the outline and image of the colony displayed in the processes of S44 to S46 in Fig. 4, the number of colonies counted, and the like, to the determination device 50 as a means for outputting information indicating the state of the colony. The environment determination system 100 can also be called an analysis system.

[0084] The DNA analysis device 40 is also called a genome analysis device, a gene analysis device, or a DNA sequencer. The DNA analysis device 40 analyzes the base sequence of the DNA of the microorganisms in the petri dish PD after the culture in the culture process is completed, and identifies the type of the microorganism. Furthermore, the DNA analysis device 40 outputs classification information of the microorganism based on the obtained type of the microorganism. The classification information is, for example, harmless, low danger level, medium danger level, high danger level, etc. Furthermore, the classification information may be, for example, a rough classification of the microorganism, such as mold, yeast, bacteria, etc.

[0085] In order to output such classification information, the DNA analyzer 40 may have a classification table that defines the correspondence between the name and symbol identification information that specify the type of each microorganism and the classification information. Such a classification table may be defined according to the business of the monitored environment. For example, the classification table may be defined according to whether the monitored environment is a biopharmaceutical research and development site, a biopharmaceutical manufacturing site, a food research and development site, a food manufacturing site, etc.

[0086] The determination device 50 performs machine learning based on teacher data in which the classification of the analysis results by the DNA analysis device 40 is added as a label to the analysis results by the analysis device 10. Machine learning is also called, for example, deep learning. The determination device 50 has a CPU, a memory, and the like, similar to the analysis device 10. As described in FIG. 1, the CPU includes a GPU, a DSP, and the like. The determination device 50 forms a multi-layered neural network between multiple virtual nodes by a computer program executablely deployed on the memory. Each node connected by the neural network is placed in one of the multiple layers. However, the neural network may be configured by a hardware circuit. For example, each node may have a CPU, a GPU, a DSP, or a product-sum operation circuit of an FPGA, and the nodes may be connected by a network. Of the multiple layers, the layer to which data from the analysis device 10 is input is the input layer. Also, of the multiple layers, the layer furthest from the input layer is the output layer.

[0087] Each node in each phase of the multiple hierarchical layers receives the output of a node in a layer close to the input layer, executes a convolution operation such as a product-sum operation or a filtering process given to itself, and passes the processing result to a node in a layer close to the output layer. The output layer receives the output of a node in a layer close to the input layer, and outputs a classification result. Here, the product-sum operation or filtering process is, for example, an inner product of a coefficient array or coefficient owned by each node and input data (vector).

[0088] In the learning stage, the labels of the correct answer data from the DNA analysis device 40 are sent together with the input data. The input data is input to the determination device 50. A set of the input data and the correct answer data is called teacher data. If the classification result, which is the output of the output layer, does not match the correct answer data, the difference value between the classification result and the correct answer data is propagated backward from the output layer to the input layer. In this backward propagation, the coefficients are adjusted in a direction in which the error at each node in each layer is reduced. By repeating such a learning stage process multiple times using different teacher data, the coefficients are adjusted and a trained model is formed. Therefore, the trained model has adjusted coefficients (coefficients of product-sum operation or filtering process) that have learned the features contained in the information such as the colony outline, image, and counted number of colonies displayed in the processes of S44 to S46 in FIG. 4. In other words, the trained model has adjusted coefficients based on the relationship between the above features and the classification information in the DNA analysis device 40 for those features.

[0089] In the recognition stage, new data from the analysis device 10 is input to the input layer. The determination device 50 executes recognition processing on the data from the analysis device 10 using the coefficients of each node in each layer adjusted in the learning stage. In other words, the trained model outputs a determination result such as classification A, B, C, etc. for each piece of input data.

[0090] Here, the input data is, for example, a set of data obtained in each culture step. For example, when the culture steps are culture step 1 to culture step FL, the input data is a polygon, image data, etc. obtained in each culture step. The polygon is, for example, ID (identification information), the number of vertices, the vertex sequence (X, Y coordinate values), the coordinates of the center of gravity, color, etc. It is assumed that the coordinate system defining the vertices of the polygon is set in advance in the analysis device 10 and the determination device 50. The input data may be either the polygon or the image data. These input data are examples of information indicating the state of a microbial colony.

[0091] The input data may also include parameters indicating environmental conditions. The parameters include temperature, humidity, and hours of sunlight. Although data for each stage of the culture process (culture processes 1 to FL) is used in Fig. 11, the processing of the determination device 50 is not limited to such processing. For example, the determination device 50 may only input data on the culture result when the final culture process FL is completed.

[0092] The determination device 50 repeatedly performs learning on the input data using teacher data to which the output result of the classification information from the DNA analysis device 40 has been added, and adjusts the coefficients of the convolution operation or filtering process. Therefore, the determination device 50 performs machine learning using teacher data including information indicating the state of the microbial colony and labels that are the classification results of the microorganisms based on the DNA base sequence obtained by analyzing the microorganisms. The CPU of the determination device 50 is an example of a means for performing machine learning.

[0093] Then, when the learning is executed a predetermined number of times and the adjustment of the coefficients progresses, the determination device 50 becomes capable of executing the processing in the recognition stage. That is, the determination device 50 outputs a classification result such as classification A, B, C, etc. for the input data from the analysis device 10.

[0094] The classification result may be, for example, Classification A = harmless, Classification B = low risk, Classification C = medium risk, Classification D = high risk, etc. The classification result may also be, for example, the type of microorganism, for example, Classification A = mold, Classification B = yeast, Classification C = bacteria, etc. The classification result may also be a more detailed classification, for example, Classification A = Actinobacteria, Classification B = Archaea, highly acidophilic bacteria, etc. The classification result may also be, for example, mold types K1, K2, K3, ... bacterial types B1, B2, B3, etc. Therefore, the determination device 50 classifies microorganisms based on information indicating the state of a newly obtained colony of microorganisms based on a trained model by machine learning. The CPU of the determination device 50 is an example of a means for classifying microorganisms and a classification device.

[0095] As described above, the environment determination system 100 of this embodiment can present the classification of microorganisms obtained in the monitored environment. That is, the environment determination system 100 performs classification based on the analysis results of the analysis device 10, such as polygonal data obtained from colonies of microorganisms, image data used in the analysis process, or data combining these with environmental parameters. Such classification may be a rough classification of the analysis results, not a strict analysis result of DNA base sequence by the DNA analysis device 40. It is sufficient if the environment determination system 100 indicates a rough classification of the collected microorganisms, so that it is possible to determine whether or not strict genome analysis is required for the collected microorganisms. By such a determination, the microorganisms collected in the monitored environment are narrowed down to a specific classification, and genome analysis is performed. As a result, the time and cost required for genome analysis are reduced.

[0096] <Third embodiment> The environment monitoring system 101 of the third embodiment will be described with reference to FIG. 12. The environment monitoring system 101 can also be called an analysis system. FIG. 12 is a diagram illustrating the environment monitoring system 101 of the third embodiment. In the above-mentioned second embodiment, a rough classification of microorganisms collected in a monitored environment was shown based on polygonal data obtained from a colony of a microorganism, image data, or data combining these with environmental parameters, which are the analysis results of the analysis device 10. In this embodiment, an environment monitoring system 101 is illustrated that monitors the state of a monitored environment by combining the analysis results or information used in the analysis by the analysis device 10 (hereinafter referred to as analysis data related to the analysis) with environmental data indicating the state of the monitored environment. As in the second embodiment, the analysis device 10 inputs information such as the outline and image of the colony displayed in the processes of S44 to S46 and the number of colonies counted to the determination device 50 as a means for outputting information indicating the state of the colony.

[0097] Strict monitoring of microorganisms is required in pharmaceutical or food research, development, or manufacturing sites. In this embodiment, a simple and highly accurate environment monitoring system 101 is described that utilizes analysis data from an analysis device 10 by appropriately selecting environmental data that indicates the state of the monitored environment.

[0098] Here, the environmental data indicating the situation in the monitored environment is, for example, the location where a sample of microorganisms is collected. The sample collection location can be divided into a location where airborne bacteria are collected and a location where attached bacteria are collected. Airborne bacteria are bacteria that float in the air or fall. When the monitored environment is inside a building, the location where airborne bacteria are collected is exemplified by a location within the building, a room number, a floor, a location within the room, etc. The location of the room is the location of the room where the petri dish PD is placed. The location of the room may be a coordinate value according to a coordinate system within the room, or identification information of a desk or furniture on which the petri dish PD is placed. However, the identification information of the desk or furniture is assumed to be fixed. The location where airborne bacteria are collected can also be a location within the monitored environment.

[0099] Adherent bacteria are collected from sample locations on the body or clothing of an operator working in the monitored environment, such as the head, hat, left and right arms, gloves, the front and back of a lab coat, the left and right sleeves of a lab coat, a mask, shoes, doorknobs, etc. Therefore, these locations are the sample locations for adhering bacteria. Furthermore, parameters indicating the situation in the monitored environment may include parameters of environmental conditions such as the number of people in the room, temperature, and humidity. These data are examples of environmental data.

[0100] A set of input data is obtained by combining the thus obtained position in the monitored environment, the sampling location, the incubation process of the petri dish PD, the polygonal data at that time (ID, number of vertices, vertex sequence, center of gravity, color), image data, the amount or number of cultivated microbial colonies, etc. Note that the amount of colonies is exemplified by the area of ​​a polygon that indicates the shape of the colonies. The polygonal data (ID, number of vertices, vertex sequence, center of gravity, color), and the data obtained in chronological order during the incubation process are The image data, the amount or number of colonies of the cultured microorganisms, etc. are examples of analysis data related to the analysis by the analysis device. In this embodiment, for example, it is assumed that the CPU 11 of the analysis device 10 or a computer (hereinafter, CPU 11, etc.) that cooperates with the analysis device 10 manages and stores the position in the monitored environment, the collection location, the culture process of the petri dish PD, etc. The CPU 11, etc., creates input data by adding information such as the collection location, the culture process of the petri dish PD, and environmental conditions to the polygon data output by the analysis device 10, the image data, the amount or number of colonies, etc., and inputs the input data to the determination device 50. The position in the monitored environment, the collection location, the culture process of the petri dish PD, the environmental conditions, etc. can be called environmental data. In addition, at that time, the determination device 50 obtains the determination result of the DNA analysis device 40 as a label of the correct answer data from the DNA analysis device 40. The determination result of the DNA analysis device 40 is, for example, OK (no corresponding microorganism detected) or NG (corresponding microorganism detected). Such a determination result is an example of a determination result for a microorganism sample after the microorganism sample is cultured. The CPU 11 or the like of the analysis device 10 may obtain the determination result from the DNA analysis device 40, label the input data with the correct answer data, and transfer the result to the determination device 50.

[0101] In this manner, also in the environment monitoring system 101, teacher data in which the input data is labeled with the correct answer data is created and input to the determination device 50. The determination device 50 executes a learning process and adjusts coefficients by inputting such teacher data. Thus, the determination device 50 executes machine learning using the teacher data. The CPU of the determination device 50 is an example of a means for executing machine learning.

[0102] Then, when the learning progresses to a certain extent and the coefficients are set to appropriate values, the determination device 50 executes the recognition process using the trained model. That is, when input data including environmental data such as the collection location, the incubation process of the petri dish PD, and environmental conditions, and analysis data including polygonal data, image data, and the amount or number of colonies, etc. are input, the determination device 50 outputs a determination result of OK or NG. Therefore, the environment monitoring system 101 can determine the status of the monitored environment using input data that combines the analysis result analysis data from the analysis device 10 and the environmental data indicating the status of the monitored environment. That is, it can be said that the determination device 50 determines the state of the monitored environment based on the newly obtained environmental data and analysis data based on the trained model by machine learning. Therefore, the CPU of the determination device 50 is an example of a determination means.

[0103] <Modification> In the above second and third embodiments, the analysis device 10 and the determination device 50 have been described as separate devices. However, the analysis device 10 and the determination device 50 may be an integrated device. For example, hardware for implementing a trained model may be incorporated in the analysis device 10 and connected to the CPU 11 via a bus. Furthermore, the CPU 11 of the analysis device 10 may construct a neural network on the main memory unit 12 and execute the learning process and the recognition process.

[0104] <Computer-readable recording medium> A program for causing a computer or other machine or device (hereinafter, referred to as a computer, etc.) to realize any of the above functions can be recorded on a recording medium readable by the computer, etc. Then, the computer, etc. can provide the function by reading and executing the program from the recording medium.

[0105] Here, a computer-readable recording medium refers to a recording medium that stores information such as data and programs electrically, magnetically, optically, mechanically, or chemically and can be read by a computer. Examples of such recording media that can be removed from a computer include flexible disks, magneto-optical disks, CD-ROMs, CD-R / Ws, DVDs, Blu-ray disks, DATs, 8mm tapes, flash drives, etc. Examples of recording media that are fixed to a computer include hard disks and ROMs (read-only memories). In addition, SSDs (Solid State Drives) are also used as recording media that can be removed from a computer. The present invention can also be used as a recording medium fixed to a computer. [Explanation of symbols]

[0106] 10 Analyzer 11 CPU 12 Main memory 13 External memory unit 14 Display section 15 Control section 16 Communications Department 17 Camera control section 18 White light control section 19 Blue light control unit 20 LED control unit 21 Fan control unit 100 Environmental Judgment System 101 Environmental Monitoring System

Claims

1. An analysis device for a collection container for collecting microorganisms, comprising: a means for acquiring a plurality of first images taken in time series using a first visible light during a culture process within a sampling container in which a sampling process of microorganisms has been performed in a monitored environment; a first identification means for detecting the occurrence of a microbial colony or the presence of a colony that has already occurred in the time series from a change in the plurality of first images in the time series, and identifying a shape of the colony whose occurrence or presence has been detected; a second identification means for identifying a shape of a microbial colony from a second image obtained by photographing the inside of the collection container with a second visible light; and a means for displaying a state of a microbial colony based on the shape identified by the first identification means and the shape identified by the second identification means.

2. The analysis device according to claim 1 , wherein the first identification means detects the emergence of a microbial colony or the presence of a colony that has already emerged based on at least one of a boundary line of a shadow present in each of the plurality of first images and a center of gravity of an area surrounded by the boundary line.

3. The analysis device according to claim 1 or 2, wherein, when a shadow that existed in an earlier image does not exist in a later image among the images before and after the time series, the first identification means excludes the shadow that existed in the earlier image as a shadow other than a microbial colony.

4. a discrimination means for discriminating between the shadows of the microbial colonies and the shadows of other than the microbial colonies; The analysis device according to claim 1 , further comprising: a presentation unit that presents information prompting an operator to intervene when it is difficult for the determination unit to make a determination.

5. The analysis device according to claim 4 , wherein the display unit displays an image captured in a state where the sampling container is illuminated with illumination light from a plurality of directions.

6. The collection container has a transparent bottom, a main body with an open top, and a transparent lid that covers the top of the main body; The analysis device includes an imaging means for capturing a first image and a second image; The analysis device according to claim 1 , further comprising a blowing means for blowing air onto a surface of a sampling container placed in a state in which the sampling container can be photographed by the photographing means.

7. The analysis device according to claim 1 , further comprising a means for receiving confirmation from a user as to whether or not a microorganism is present based on the displayed state of the colonies.

8. 8. The analysis device according to claim 1, further comprising: a means for setting, based on a plurality of images captured by gradually changing the amount of light using one of the first visible light and the second visible light, an amount of light at a stage where a change in the number of shadows among a plurality of images captured by gradually changing the amount of light satisfies a predetermined condition, to an amount of light when capturing images using the first visible light or the second visible light.

9. The analytical device according to claim 1 , further comprising a means for displaying the state of the colonies in chronological order.

10. a means for performing machine learning using training data including information indicating the state of the colony of the microorganism and a label that is a classification result of the microorganism based on a DNA base sequence obtained by analyzing the microorganism; Based on the machine learning trained model, the newly obtained colony of the microorganism is The analysis device according to claim 1 , further comprising: a means for classifying the microorganisms based on information indicating a state.

11. a means for performing machine learning using training data including environmental data including at least one of a position in the monitored environment where the microbial sample was collected by the collection container, a collection target location, a stage in the culture process where the sample was cultured, and environmental conditions in the monitored environment, analysis data related to analysis by the analysis device, and a label which is a judgment result for the microbial sample after the microbial sample has been cultured; and The analysis device described in any one of claims 1 to 9, further comprising a means for determining the state of the monitored environment based on the newly obtained environmental data and the analysis data based on the learned model by machine learning.

12. an analyzer for sampling a collection vessel for sampling microorganisms; A classification device, The analysis device comprises: a means for acquiring a plurality of first images taken in time series using a first visible light during a culture process within a sampling container in which a sampling process of microorganisms has been performed in a monitored environment; a first identification means for detecting the occurrence of a microbial colony or the presence of a colony that has already occurred in the time series from a change in the plurality of first images in the time series, and identifying a shape of the colony whose occurrence or presence has been detected; a second identification means for identifying a shape of a microbial colony from a second image obtained by photographing the inside of the collection container with a second visible light; and a means for outputting information indicating a state of a microbial colony based on the shape identified by the first identification means and the shape identified by the second identification means, The classification device comprises: a means for performing machine learning using training data including information indicating the state of the colony of the microorganism and a label that is a classification result of the microorganism based on a DNA base sequence obtained by analyzing the microorganism; A means for classifying the microorganisms based on newly obtained information indicating the state of the colonies of the microorganisms based on the trained model by machine learning; An analysis system comprising:

13. an analyzer for sampling a collection vessel for sampling microorganisms; A determination device, The analysis device comprises: a means for acquiring a plurality of first images taken in time series using a first visible light during a culture process within a sampling container in which a sampling process of microorganisms has been performed in a monitored environment; a first identification means for detecting the occurrence of a microbial colony or the presence of a colony that has already occurred in the time series from a change in the plurality of first images in the time series, and identifying a shape of the colony whose occurrence or presence has been detected; a second identification means for identifying a shape of a microbial colony from a second image obtained by photographing the inside of the collection container with a second visible light; and a means for outputting information indicating a state of a microbial colony based on the shape identified by the first identification means and the shape identified by the second identification means, The determination device includes: a means for performing machine learning using training data including environmental data including at least one of a position in the monitored environment where the microbial sample was collected by the collection container, a collection target location, a stage in the culture process where the sample was cultured, and environmental conditions in the monitored environment, analysis data related to analysis by the analysis device, and a label which is a judgment result for the microbial sample after the microbial sample has been cultured; and A means for determining a state of the monitored environment based on the newly obtained environmental data and the analysis data, based on the trained model by the machine learning; An analysis system comprising:

14. An analysis method in which a computer analyzes a collection container for collecting microorganisms, comprising: acquiring a plurality of first images taken in time series using a first visible light in a culture process inside a sampling container in which a sampling process of microorganisms has been performed in a monitored environment; A first step of detecting the occurrence of a colony of a microorganism or the presence of a colony that has already occurred in the time series from a change in the plurality of first images in the time series, and identifying a shape of the colony whose occurrence or presence has been detected; a second step of identifying a shape of a microbial colony from a second image captured inside the collection container using a second visible light; and displaying a state of the microbial colony based on the shape identified in the first step and the shape identified in the second step.

15. A program for causing a computer to analyze a collection container for collecting microorganisms, acquiring a plurality of first images taken in time series using a first visible light in a culture process inside a sampling container in which a sampling process of microorganisms has been performed in a monitored environment; A first step of detecting the occurrence of a colony of a microorganism or the presence of a colony that has already occurred in the time series from a change in the plurality of first images in the time series, and identifying a shape of the colony whose occurrence or presence has been detected; a second step of identifying a shape of a microbial colony from a second image captured inside the collection container using a second visible light; and a step of displaying a state of a microbial colony based on the shape specified in the first step and the shape specified in the second step.

Images