Inspection device

Abstract

A water quality determination device 30 receives an input of a test strip image obtained by photographing a water quality test strip and the illuminance of the photographing place, acquires a color code for the reaction area of the water quality test strip from the test strip image, corrects the color code to the color code at a reference illuminance according to the illuminance, and determines the water quality parameter on the basis of the corrected color code.

Description

Inspection Equipment

[0001] The present disclosure relates to an inspection apparatus.

[0002] Demand for land-based aquaculture is on the rise amid concerns about rising sea temperatures due to global warming and the resulting impact on global protein supplies. Furthermore, given the current state of serious marine pollution, consumers are increasingly calling for safe, virus-free seafood, and land-based aquaculture is gaining attention as one solution. Indeed, the land-based aquaculture market is experiencing vigorous growth, with new entrants from a variety of industries.

[0003] In aquaculture, NO causes oxygen deficiency and respiratory distress in fish. - 2 (nitrites), which can cause serious illnesses - Water quality testing to detect nitrate levels is essential. Water quality testing is often done manually using water quality test strips that change color due to a chemical reaction.

[0004] Kazuki Fukae, "Research on Optimal Feeding Systems in Aquaculture Using IoT," Nagasaki University, July 2023, Internet <URL: https: / / nagasaki-u.repo.nii.ac.jp / record / 2000237 / files / KOK131_Fukae.pdf>, "Water Quality in Fish Farming," T&C Technical Co., Ltd., Internet <URL: http: / / www.tactec.jp / download / reference / ff10182-a_water_quality_management_in_aquaculture.pdf>, Ryohei Yasutomi and Kazufumi Imada, "Water Quality Analysis Items Used to Determine the Suitability of Fish Farming Water," Fish and Water, Hokkaido Research Organization, 2012, 49-1, pp. 13-22, "Efficient Growth Environment Management in Land-Based Aquaculture," Tochigi Prefecture, Internet < URL: https: / / www.pref.tochigi.lg.jp / f01 / documents / 2-4.pdf〉

[0005] Nitrogen compounds such as ammonia generated from leftover feed and excrement are harmful to fish health and can cause fish diseases. Daily inspections are required at aquaculture sites, and a practical sensor that is suited to on-site operations is desired.

[0006] Current water quality testing uses test strips that identify chemical reactions by their hue, but these strips are easily affected by light. For example, on sunny days, the test strips appear bright due to strong sunlight, while on cloudy days they appear dark. Because the test results are based on human visual judgment, there is variation and misjudgment depending on the influence of ambient light.

[0007] The present disclosure has been made in view of the above, and aims to provide a water quality evaluation technology that can obtain stable and highly immediate test results.

[0008] An inspection device according to one aspect of the present disclosure includes an input unit that inputs an image of a test paper and the illuminance at the location where the image was taken, a correction unit that acquires pixel values ​​of the reaction area of ​​the test paper from the image and corrects the pixel values ​​to pixel values ​​at a standard illuminance in accordance with the illuminance, and a determination unit that determines inspection parameters based on the corrected pixel values.

[0009] According to the present disclosure, a water quality evaluation technique that can obtain stable and highly immediate test results can be provided.

[0010] FIG. 1 is a diagram showing an example of the configuration of a water quality testing system according to this embodiment. FIG. 2 is a diagram showing an example of changes in color codes according to illuminance registered in a color code database. FIG. 3 is a diagram showing an example of a color chart for water quality test strips. FIG. 4 is a sequence diagram showing an example of the processing flow of the water quality testing system. FIG. 5 is a diagram showing an example of an operation screen when photographing a water quality test strip. FIG. 6 is a diagram showing an example of the hardware configuration of a water quality determination device.

[0011] An example of the configuration of a water quality testing system according to this embodiment will be described with reference to Figure 1. The water quality testing system shown in the figure includes a terminal 10 and a water quality determination device 30. The terminal 10 and the water quality determination device 30 are connected via a network.

[0012] The user measures the water quality using a water test strip, photographs the water test strip, and measures the illuminance using the terminal 10. When the terminal 10 sends the photographed image of the water test strip (hereinafter referred to as the "test strip image") and the illuminance to the water quality assessment device 30, the water quality assessment result is returned from the water quality assessment device 30 and displayed on the terminal 10.

[0013] Water quality test strips can measure specific water quality parameters by immersing the test strip in water and comparing the change in hue of the reaction area (also called the test area or detection area) with a color chart. The reaction area refers to a specific part of the test strip that changes color depending on the concentration or presence of a substance to be detected. Some water quality test strips have multiple reaction areas that can determine multiple parameters, but this section will focus on one parameter. To determine multiple parameters, simply perform the same process for each parameter. The terminal 10 and the water quality determination device 30 will be described below.

[0014] The terminal 10 includes a photographing unit 11 for photographing the water quality test strip, an illuminance measuring unit 12 for measuring illuminance, a communication unit 13 for communicating with the water quality assessment device 30, and a user interface 14 for accepting user operations and displaying the water quality assessment results. For example, a smartphone can be used as the terminal 10. An existing camera app running on the smartphone is used to photograph the water quality test strip, an illuminance meter app is used to measure the illuminance, and the test strip image and illuminance are then uploaded to the water quality assessment device 30. Alternatively, an application can be developed to perform a series of processes, including photographing the water quality test strip, measuring the illuminance, uploading the test strip image and illuminance, and displaying the water quality assessment results, and run on the smartphone. The terminal 10 is not limited to a smartphone; a PC or other mobile device can also be used.

[0015] The water quality determination device 30 includes a correction unit 31 , a color code database 32 , a water quality determination unit 33 , a water quality database 34 , and a communication unit 35 .

[0016] The correction unit 31 obtains the color code of the pixels in the reaction area from the test strip image and, by referring to the color code database 32, corrects the obtained color code according to the illuminance. A color code is a code in which the value of each RGB color is expressed as a six-digit hexadecimal number followed by a sharp (#). Illuminance represents the amount of light shining from a light source onto the water quality test strip, and is measured in lux (lx). The illuminance is received from the terminal 10 along with the test strip image.

[0017] The color code database 32 stores information about changes in color codes due to illuminance. FIG. 2 shows an example of changes in color codes due to illuminance stored in the color code database 32. FIG. 2 shows an example of how each color sample (the color code after change) appears when a color chart having six color samples, as shown in FIG. 3, is photographed at seven levels of illuminance from 0 lux to 1000 lux. In other words, FIG. 2 can also be said to show the changes in six color samples (color codes) at each illuminance. The number of color samples is not limited to six. The changes in color codes due to illuminance shown in FIG. 2 are stored in the color code database 32 as linked data sets. Data sets for each water quality parameter (nitrite concentration, nitrate concentration, etc.) are registered in the color code database 32.

[0018] The correction unit 31 determines a search location in the color code database 32 using the received illuminance, and searches for a color sample similar to the color code obtained from the test paper image from that search location. For example, if the received illuminance is 46 lux, the correction unit 31 searches for a color code from a column of color samples obtained within the range of 46±10 lux. In the example of Figure 2, the correction unit 31 searches for a color sample similar to the color code from the column of 50 lux.

[0019] If the search location contains multiple color sample columns, the correction unit 31 searches for a color code from among the multiple color sample columns. For example, if the received illuminance is 46 lux and a 40 lux column and a 50 lux column are registered in the color code database 32, the correction unit 31 searches for a color sample that is close to the color code from the 40 lux column and the 50 lux column.

[0020] The correction unit 31 corrects the color code of the test paper image based on the difference between the color sample obtained by the search (hereinafter referred to as color sample 1) and a color sample (hereinafter referred to as color sample 2) located in the same position as color sample 1 at a standard illuminance (e.g., 300 lux). The standard illuminance can be set arbitrarily. For example, if the color code of the test paper image is closest to the second-lowest color sample (color sample 1) in the 50 lux column, the correction unit 31 corrects the color code of the test paper image based on the difference between color sample 1 and the second-lowest color sample (color sample 2) in the 300 lux column.

[0021] When searching for a color code, the correction unit 31 converts the color code to HSV and searches the search location column for a color sample with a similar saturation S and value V. The HSV model consists of three components: hue, saturation, and value. Conversion from RGB (red, green, blue) to HSV can be performed using the following formula:

[0022]

[0023] Here, MAX in the formula is the largest value among RGB, and MIX is the smallest value among RGB.

[0024] The correction unit 31 corrects the color code of the test paper image using the difference in saturation S and brightness V between the color sample 1 obtained by the search and the color sample 2 with the standard illuminance.

[0025] The inventors immersed water quality test strips in chemicals of various concentrations, photographed them together with a color chart, obtained color data (RGB and HSV), and performed multiple regression analysis on the obtained data. As a result of the analysis, they narrowed down the factors that affect the concentration of the test strips to G, S, and V. Since G is an element that indicates color, and saturation S and value V are indicators of color vividness and brightness, in this embodiment, saturation S and value V are used to correct for changes in illuminance. Note that search and correction may also be performed using the RGB values ​​of the color code.

[0026] If the color code corresponding to the illuminance received from the terminal 10 is not stored in the color code database 32, the correction unit 31 inputs the illuminance into a machine learning model that has learned the relationship between color codes and illuminance, and estimates the color code for that illuminance. For example, if the received illuminance is 88 lux, in the example of FIG. 2 , the color code database 32 does not store a column of color samples within the range of 88±10 lux. Therefore, the correction unit 31 estimates a column of color samples within the range of 88±10 lux (e.g., a column for 80 lux, or a column for 88 lux or 90 lux). The estimated column for 80 lux is added to the color code database 32, and a dataset is constructed. The reconstructed dataset is stored as is in the color code database 32 and is used when new data is received from the terminal 10.

[0027] As another method of correcting the color code, the correction unit 31 may input the color code and illuminance obtained from the test paper image into a machine learning model that has learned the relationship between the color code and illuminance by machine learning, and obtain a corrected color code at a standard illuminance.

[0028] The water quality determination unit 33 references the water quality database 34 and determines the water quality based on the corrected color code. The water quality database 34 stores changes in color codes corresponding to water quality parameters at a standard illuminance. For example, the water quality database 34 stores the concentrations (the numerical values ​​(mg / L) on the right side of FIG. 3 ) of nitrite and nitrate corresponding to each color sample at a standard illuminance (e.g., 300 lux) for each of them. The water quality determination unit 33 searches for a color sample that is close to the corrected color code and outputs the concentration corresponding to that color sample.

[0029] Alternatively, the water quality determination unit 33 may calculate the concentration using the RGB values ​​of the corrected color code. For this calculation, an equation based on multiple regression analysis using the RGB values ​​as explanatory variables can be used.

[0030]

[0031] Here, a1, a2, and a3 are weights for calculating the concentration with high accuracy. Since the G value is useful as an explanatory variable, it is preferable to set a2>a1, a3.

[0032] The communication unit 35 communicates with the terminal 10 via the network. For example, the communication unit 35 receives a test strip image and illuminance from the terminal 10 and transmits the water quality evaluation result to the terminal 10.

[0033] An example of the processing flow of the water quality testing system of this embodiment will be described with reference to the sequence diagram of Figure 4. It is assumed that the test results using the water quality test strips (water quality test strips after reaction) have already been obtained.

[0034] In step S11, the user operates the terminal 10 to take a photograph of the water quality test paper after the reaction.

[0035] For example, as shown in the operation screen 100 of Figure 5, the terminal 10 displays a marker 110 on the operation screen 100. The user operates the terminal 10 so that the reaction area 210 of the water quality test strip 200 fits within the marker 110, and then photographs the water quality test strip 200. If the marker 110 is not displayed, the user may photograph the water quality test strip 200 so that the reaction area 210 is near the center of the operation screen 100. If the terminal 10 or the water quality determination device 30 can detect the reaction area 210 from the test strip image, the user may photograph the water quality test strip 200 so that the reaction area 210 fits within the screen.

[0036] In step S12, the user measures the illuminance by operating the terminal 10. The illuminance measurement may be performed at the same time as step S11 or before step S11. The user may measure the illuminance using an illuminance meter other than the terminal 10 and input the measured illuminance into the terminal 10.

[0037] In step S13, the terminal 10 uploads the test strip image and the illuminance to the water quality determination device 30. The terminal 10 may also cut out an image of the reaction area 210 of the test strip image and upload it to the water quality determination device 30. When uploading the test strip image and illuminance to the water quality determination device 30, the terminal 10 may also send the environmental conditions to the water quality determination device 30. The environmental conditions are the conditions of the environment in which the water quality test strip was photographed, such as the location type (indoor or outdoor) and whether or not the lights were on.

[0038] In step S14, the water quality determining device 30 obtains a color code of the image of the reaction area 210 from the test strip image, and corrects the obtained color code according to the illuminance.

[0039] In step S15, the water quality determining device 30 determines the water quality based on the corrected color code.

[0040] In step S16 , the water quality determination device 30 transmits the water quality determination result to the terminal 10 .

[0041] The terminal 10 displays the water quality assessment results received from the water quality assessment device 30. The terminal 10 may display the water quality parameters as numerical values, may display the state of the water quality (good or bad) based on the water quality parameters, or may display the color code and color used in the assessment.

[0042] As explained above, the water quality determination device 30 of this embodiment inputs a photographed test strip image and the illuminance of the location where the image was taken, obtains a color code of the reaction area of ​​the water quality test strip from the test strip image, corrects the color code to a color code at a standard illuminance according to the illuminance, and determines water quality parameters based on the corrected color code. This allows for stable and immediate water quality test results to be obtained regardless of the illuminance conditions of the environment in which the water quality test is conducted.

[0043] Although water quality testing has been described in this disclosure, the technology of this disclosure can also be applied to tests other than water quality testing that measure parameters based on color changes in test strips, such as formaldehyde detection test strips that detect components in gases by color, urine tests, allergy patch tests, pH test strips, and blood tests.

[0044] The water quality determining device 30 described above can be, for example, a general-purpose computer system including a central processing unit (CPU) 901, memory 902, storage 903, communication device 904, input device 905, and output device 906, as shown in Fig. 6. In this computer system, the water quality determining device 30 is realized by the CPU 901 executing a predetermined program loaded onto the memory 902. This program can be recorded on a computer-readable non-transitory recording medium such as a magnetic disk, optical disk, or semiconductor memory, or can be distributed via a network.

[0045] REFERENCE SIGNS LIST 10 Terminal 11 Photography unit 12 Illuminance measurement unit 13 Communication unit 14 User interface 30 Water quality determination device 31 Correction unit 32 Color code database 33 Water quality determination unit 34 Water quality database 35 Communication unit

Claims

1. An inspection device comprising: an input unit that inputs an image of a test paper and the illuminance at the location where the image was taken; a correction unit that acquires pixel values ​​of the reaction area of ​​the test paper from the image and corrects the pixel values ​​in accordance with the illuminance; and a determination unit that determines inspection parameters based on the corrected pixel values.

2. An inspection device according to claim 1, comprising a color code database that holds a sequence of color samples for each illuminance, and wherein the correction unit searches for a color sample close to the pixel value from the sequence of color samples corresponding to the input illuminance, and corrects the pixel value based on the difference between the searched color sample and a reference color sample.

3. An inspection device according to claim 2, wherein the correction unit determines the saturation and brightness of the pixel value, and searches the sequence of color samples for a color sample having similar saturation and brightness.

4. An inspection device according to claim 2, wherein, if the color code database does not hold a sequence of color samples corresponding to the input illuminance, the correction unit inputs the illuminance into a machine learning model that has learned the relationship between illuminance and color samples by machine learning, estimates the sequence of color samples for that illuminance, and registers the estimated sequence of color samples in the color code database.

Images