Apparatus for predicting pollution level of hull based on machine learning for managing hull and method thereof

Abstract

The present invention relates to a machine learning-based hull bottom pollution prediction device and method for hull bottom management. According to an embodiment of the present invention, a machine learning-based hull bottom pollution prediction device for hull bottom management comprises: an input unit that receives evaluation target variable data for each vessel regarding a plurality of factors affecting hull bottom pollution from a user; a data preprocessing unit that standardizes the evaluation target variable data; and a prediction unit that predicts hull bottom pollution by applying the preprocessed variable data to a pre-trained learning model. As such, according to the present invention, the degree of contamination of the hull bottom can be predicted without a hull bottom inspection by learning variable data based on a plurality of factors affecting the degree of contamination of the hull bottom. Furthermore, unnecessary costs for hull bottom cleaning or hull bottom inspection can be reduced by utilizing the predicted degree of contamination of the hull bottom. In addition, it has the effect of improving the reliability of the predicted bottom contamination level by calculating the similarity between variable data based on multiple factors affecting bottom contamination.

Description

Technology Field

[0001] The present invention relates to a machine learning-based bottom contamination prediction device and method for bottom management, and more specifically, to a bottom contamination prediction device and method that predicts the contamination level of the bottom of a ship by applying variable data to artificial intelligence. Background Technology

[0002] Contaminants such as shellfish and water stains accumulating on the hull cause problems such as reduced vessel speed or increased fuel consumption. Therefore, periodic hull inspections are performed for hull maintenance.

[0003] Generally, hull inspections involve photographing the bottom of a moored vessel using divers or underwater drones, assessing the degree of contamination based on the captured footage and images, and determining whether to perform hull cleaning accordingly. However, since criteria for assessing contamination vary depending on the observer even when viewing the same footage, disputes arise when determining the level of contamination; in particular, this results in unnecessary costs for hull cleaning and inspection due to the subjective judgment of the shipowner.

[0004] Therefore, there is a need for a technology that can predict the degree of hull contamination without hull inspection by utilizing variable data related to hull contamination.

[0005] The technology forming the background of the present invention is disclosed in Korean Registered Patent Publication No. 10-1454855 (published on October 28, 2014). The problem to be solved

[0006] The present invention aims to provide a hull bottom contamination prediction device and method for predicting the contamination level of a hull bottom without a hull bottom inspection by learning variable data based on a plurality of factors affecting the contamination level of a hull bottom. means of solving the problem

[0007] According to an embodiment of the present invention for achieving such technical challenges, a machine learning-based bottom pollution prediction device for bottom management comprises: an input unit that receives evaluation target variable data for each vessel regarding a plurality of factors affecting bottom pollution from a user; a data preprocessing unit that standardizes the evaluation target variable data; and a prediction unit that predicts bottom pollution by applying the preprocessed variable data to a pre-trained learning model.

[0008] Factors affecting the degree of bottom contamination may include at least one of the following: the start time of bottom contamination, the date of bottom inspection, the number of bottom inspections, the time of mooring of the vessel at the quay, the presence or absence of a typhoon during the mooring period, the mooring period, the mooring quay, the specific gravity of seawater, the seawater temperature, whether cleaning was performed prior to the bottom inspection, the manufacturer of the bottom plating paint, the type of antifouling paint on the vessel's plating, and the antifouling performance of each paint.

[0009] The above preprocessing unit can calculate the Pearson correlation coefficient for each input variable data subject to evaluation, calculate the variance expansion index to remove factors with high correlation between factors affecting bottom contamination, and standardize the multiple variable data subject to evaluation into a certain unit.

[0010] It may further include a learning unit that sets the above-mentioned learning target variable data as input data and sets the bottom contamination level as output data to train a learning model, a calculation unit that calculates the similarity between the above-mentioned evaluation target variable data and the learning target variable data, an output unit that outputs the above-mentioned predicted bottom contamination level, and a classification unit that classifies the evaluation target variable data according to the above-mentioned predicted contamination level.

[0011] The above output unit may further output at least one of the variable data of the training target data most similar to the evaluation target data for which prediction has been completed and the actual contamination value, depending on whether a user requests it.

[0012] The above prediction unit may assign a high weight to mooring periods by quay, monthly mooring periods, seasons, or antifouling performance among the factors affecting the hull bottom contamination, and assign a low weight to water temperature or specific gravity.

[0013] It may further include a data augmentation unit that augments the distribution of the learning target data so that the distribution of the above-classified learning target data is similar to a normal distribution.

[0014] According to another embodiment of the present invention, a method for predicting bottom pollution using a machine learning-based bottom pollution prediction device for bottom management comprises the steps of: receiving evaluation target variable data for each vessel regarding a plurality of factors affecting bottom pollution from a user; standardizing the evaluation target variable data; and applying the preprocessed variable data to a pre-trained learning model to predict bottom pollution. Effects of the invention

[0015] As such, according to the present invention, the degree of contamination of the hull bottom can be predicted without a hull bottom inspection by learning variable data based on a plurality of factors affecting the degree of contamination of the hull bottom. Furthermore, unnecessary costs for hull bottom cleaning or hull bottom inspection can be reduced by utilizing the predicted degree of contamination of the hull bottom.

[0016] In addition, it has the effect of improving the reliability of the predicted bottom contamination level by calculating the similarity between variable data based on multiple factors affecting bottom contamination. Brief explanation of the drawing

[0017] FIG. 1 is a configuration diagram of a machine learning-based bottom contamination prediction device for bottom management according to an embodiment of the present invention. FIGS. 2a and FIGS. 2b are illustrative diagrams for explaining variable data according to an embodiment of the present invention. FIG. 3a is an illustrative diagram explaining the effect of mooring period by quay on bottom contamination according to an embodiment of the present invention. FIG. 3b is an illustrative diagram explaining the effect of seawater specific gravity on bottom contamination according to an embodiment of the present invention. FIG. 3c is an illustrative diagram explaining the effect of seawater temperature on bottom contamination according to an embodiment of the present invention. FIG. 3d is an illustrative diagram explaining the effect of antifouling performance according to an embodiment of the present invention on bottom contamination. FIGS. 4a and FIGS. 4b are illustrative diagrams for explaining variable data preprocessing according to an embodiment of the present invention. FIG. 5 is a flowchart of a machine learning-based method for predicting bottom contamination according to an embodiment of the present invention. FIG. 6 is an illustrative diagram for explaining the degree of bottom contamination according to an embodiment of the present invention. Figure 7 is an example diagram illustrating step S540 of Figure 5. Figure 8 is an example diagram illustrating step S560 of Figure 5. Specific details for implementing the invention

[0018] Then, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification have been given similar reference numerals.

[0019] Throughout the specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected," but also cases where they are "electrically connected" with other components interposed between them. Furthermore, when a part is described as "including" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0020] Then, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the present invention.

[0021] Below, we will examine a bottom contamination prediction device according to an embodiment of the present invention through FIGS. 1 to 4b.

[0022] FIG. 1 is a configuration diagram of a machine learning-based bottom contamination prediction device for bottom management according to an embodiment of the present invention.

[0023] As shown in FIG. 1, the bottom pollution prediction device (100) according to an embodiment of the present invention includes an input unit (110), a data preprocessing unit (120), and a prediction unit (130), and may further include a learning unit (140), a classification unit (150), a calculation unit (160), an output unit (170), and a data augmentation unit (180).

[0024] First, the input unit (110) receives evaluation target variable data for each vessel regarding multiple factors affecting the degree of bottom contamination from the user.

[0025] FIGS. 2a and 2b are illustrative diagrams for explaining variable data subject to evaluation according to an embodiment of the present invention.

[0026] The variable data subject to evaluation are the variable data values ​​based on factors affecting the degree of bottom contamination for the vessel to be evaluated.

[0027] As shown in FIGS. 2a and 2b, factors affecting the degree of bottom contamination may include at least one of the following: the start time of bottom contamination, the date of bottom inspection, the number of bottom inspections, the time of mooring of the vessel at the quay, the presence or absence of a typhoon during the mooring period, the mooring period, the mooring quay, the specific gravity of seawater, the seawater temperature, whether cleaning was performed prior to the bottom inspection, the manufacturer of the bottom plating paint, the type of antifouling paint on the vessel's plating, and the antifouling performance of each paint.

[0028] And the data preprocessing unit (120) calculates the Pearson correlation coefficient for each factor affecting the bottom of the hull contamination and calculates the variance expansion factor (VIF) to remove factors with a large correlation between factors affecting the bottom of the hull contamination.

[0029] Therefore, the data preprocessing unit (120) can accurately predict the pollution level by removing elements with high correlation and using only the major factors that affect the pollution level of the ship's bottom.

[0030] FIG. 3a is an illustrative diagram illustrating the effect of mooring period by quay on bottom contamination according to an embodiment of the present invention, FIG. 3b is an illustrative diagram illustrating the effect of seawater specific gravity on bottom contamination, FIG. 3c is an illustrative diagram illustrating the effect of seawater temperature on bottom contamination, and FIG. 3d is an illustrative diagram illustrating the effect of antifouling performance on bottom contamination.

[0031] For example, as shown in Fig. 3a, the Pearson correlation coefficient for the mooring periods at quays 1 and 5 is 0.27, and it is assumed that the longer the mooring period at quays 1 and 5, the more likely the contamination level is to increase. Also, the Pearson correlation coefficient for the mooring periods at quays 2 to 4 is -0.13, and it is assumed that the longer the mooring period at quays 2 to 4, the more likely the contamination level is to decrease.

[0032] Therefore, the data preprocessing unit (120) can determine that the period of mooring at quays 1 and 5 is a more important factor than the period of mooring at quays 2 to 4 in predicting the contamination level of the hull bottom.

[0033] As shown in Figure 3b, as the specific gravity of seawater increases, the degree of hull contamination tends to decrease.

[0034] The specific gravity of seawater according to an embodiment of the present invention is the average value of the specific gravity of seawater during the mooring period.

[0035] As shown in Fig. 3c, the degree of hull contamination tends to increase as the seawater temperature increases.

[0036] The seawater temperature according to an embodiment of the present invention is the average value of the seawater temperature during the mooring period.

[0037] As shown in Fig. 3d, as the antifouling performance increases, the bottom contamination tends to decrease.

[0038] The antifouling performance according to an embodiment of the present invention can be classified into one of "high," "medium-high," "medium," "medium-low," "low," and "lowest" depending on the antifouling performance of each paint, and can be changed according to the user's settings.

[0039] And the data preprocessing unit (120) standardizes the unit of the input variable data of the element.

[0040] FIGS. 4a and FIGS. 4b are illustrative diagrams for explaining variable data preprocessing according to an embodiment of the present invention.

[0041] For example, it is assumed that the bottom contamination prediction device (100) receives input from the user for the monthly mooring period (period 7-11), the quay mooring period (period 15Q and period 234Q) and the antifouling performance (paint_performance_final) for each vessel, as shown in FIG. 4a.

[0042] As shown in Figure 4a, since variable data have different units depending on the factors affecting bottom contamination, errors may occur when predicting contamination.

[0043] Accordingly, as shown in FIG. 4b, the data preprocessing unit (120) converts the units between variable data according to factors affecting the bottom contamination level to a constant level.

[0044] And the prediction unit (130) applies the preprocessed variable data to a pre-trained learning model to predict the contamination level of the ship's bottom.

[0045] In general, mooring periods by quay, monthly mooring periods, seasons, or antifouling performance have a greater influence on bottom contamination. Conversely, water temperature or specific gravity has less influence on bottom contamination compared to mooring periods by quay, monthly mooring periods, seasons, or antifouling performance.

[0046] Therefore, the prediction unit (130) may give a higher weight to the mooring period by quay, the mooring period by month, the season, or the antifouling performance among the factors affecting the bottom of the hull, and give a lower weight to the water temperature or specific gravity.

[0047] Then, the learning unit (140) sets the variable data to be learned as input data and sets the hull contamination level as output data to train the learning model.

[0048] And the classification unit (150) classifies the learning target variable data according to the pollution level of the output hull bottom.

[0049] The classification unit (150) according to an embodiment of the present invention can classify learning target variable data in 0.5 step increments from 0 to 4 steps according to the output pollution level value, and can be changed according to the user's settings.

[0050] And the operation unit (160) calculates the similarity between the variable data to be evaluated and the variable data to be learned.

[0051] The hull bottom contamination prediction device (100) can improve the reliability of the hull bottom contamination by outputting a similarity and providing it to the user.

[0052] And the output unit (170) outputs the predicted contamination level of the ship's bottom.

[0053] The output unit (170) can output at least one of the variable data of the training target data most similar to the evaluation target data for which prediction has been completed and the actual contamination value, depending on whether the user requests it.

[0054] Data of multiple target variables classified according to contamination levels is not suitable for use as training data because the distribution by contamination level is not uniform.

[0055] And the data amplification unit (180) amplifies the classified learning target variable data.

[0056] The data amplification unit (180) can amplify the distribution of the target variable data into a form similar to a normal distribution using the SMOGN (Synthetic Minority Over-Sampling Technique for Regression with Gaussian Noise) method.

[0057] Hereinafter, a method for predicting bottom contamination using a machine learning-based bottom contamination prediction device for bottom management according to an embodiment of the present invention will be described through FIGS. 5 to 8.

[0058] FIG. 5 is a flowchart of a machine learning-based method for predicting bottom contamination according to an embodiment of the present invention, and FIG. 6 is an illustrative diagram for explaining bottom contamination according to an embodiment of the present invention.

[0059] First, as shown in FIG. 5, the input unit (110) receives evaluation target variable data for each of the multiple factors affecting the hull bottom contamination level for each vessel (S510).

[0060] In this case, the antifouling performance of each paint can be classified based on the contamination levels between the hull bottom and the propeller. If there is a difference of 3 levels or more in contamination levels between the hull bottom and the propeller, or if the contamination level of the hull bottom is less than 0.5 regardless of the contamination level of the propeller, the antifouling performance of each paint can be classified as "High." If there is a difference of 2 levels or more in contamination levels between the hull bottom and the propeller, or if the contamination level of the hull bottom is greater than 0.5 and less than 1 regardless of the contamination level of the propeller, the antifouling performance of each paint can be classified as "Medium-High." If there is a difference of 1 level or more in contamination levels between the hull bottom and the propeller, the antifouling performance of each paint can be classified as "Medium-Low," and if the contamination levels between the hull bottom and the propeller are the same, it can be classified as "Low." Furthermore, if the contamination level of the hull bottom is higher than the contamination level of the propeller, the antifouling performance of each paint can be classified as "Lowest."

[0061] As seen in Figure 6, the contamination of the hull or propeller is classified into stages based on the presence and distribution area of ​​barnacles or water stains on the image.

[0062] The hull bottom contamination prediction device (100) can determine the partial contamination level as level 0 if there are no barnacles or water stains on the stern, mid, and bow images or propeller images, and can determine level 1 if there are no barnacles and only water stains. The judgment unit (120) can determine level 2 if the distribution of barnacles on the stern, mid, and bow images or propeller images is discontinuous and 25% or less, level 3 if the distribution of barnacles is continuous and greater than 25% and less than 40%, and level 4 if the outer plating is contaminated by 40% or more.

[0063] And the data preprocessing unit (120) standardizes the variable data to be evaluated (S520).

[0064] The data preprocessing unit (120) can standardize the unit of variable data consistently by applying the following mathematical formula 1.

[0065]

[0066] And the prediction unit (130) applies the variable data to the learning model that has been trained to predict the contamination level of the ship's bottom (S530).

[0067] At this time, the learning model is trained using a 5-fold cross-validation method, and the prediction unit (130) can predict the contamination level of the ship's bottom using the learning model.

[0068] Next, the output unit (170) outputs the predicted contamination value of the hull bottom in step S530 (S540).

[0069] Figure 7 is an example diagram illustrating step S540 of Figure 5.

[0070] As shown in Fig. 7, the user can simply determine whether cleaning is necessary based on the output contamination level. The user can determine that cleaning is not necessary when the contamination level is 0 to 2, determine whether cleaning is necessary based on the degree of contamination when it is 3, and determine that cleaning is necessary when it is 4.

[0071] When a request is received from a user to output similar data for the predicted variable data (S550), the operation unit (160) calculates the similarity between the predicted variable data and the training target variable data (S560).

[0072] Figure 8 is an example diagram illustrating step S560 of Figure 5.

[0073] The operation unit (160) applies the predicted evaluation target variable data and the learning target variable data to the previously trained learning model and cosine similarity, and calculates the similarity between the evaluation target variable data and the learning target variable data as shown in FIG. 8.

[0074] At this time, the learning model can be trained using at least one of Random Forest, Gradient Boosting, SVM (Support Vector Machines), LightBGM, XGBoost, and GaussianProcessRegressor.

[0075] Next, the bottom pollution prediction device (100) selects the learning target variable data that is most similar to the evaluation target variable data that has completed prediction using the similarity calculated in step S560 (S570).

[0076] For example, assume that the similarity between the learning target variable data and the evaluation target variable data for each vessel is calculated as 90% for vessel 1234, 83% for vessel 2277, 78% for vessel 2676, and 75% for vessel 8178. Then, the hull bottom contamination prediction device (100) selects the variable data of vessel 1234 as the learning target variable data most similar to the evaluation target variable data.

[0077] The output unit (170) outputs at least one of the similarity value of the training target variable data most similar to the evaluation target variable data for which prediction is complete and the actual contamination value of the most similar variable data (S580).

[0078] The bottom pollution prediction device (100) can improve the reliability of the predicted pollution level by providing the user with the similarity of the learning target variable data most similar to the evaluation target variable data and the actual pollution level value of the most similar variable data.

[0079] As such, according to an embodiment of the present invention, the contamination level of the hull bottom can be predicted without a hull bottom inspection by learning variable data based on a plurality of factors affecting the contamination level of the hull bottom. Furthermore, unnecessary hull bottom cleaning costs or hull bottom inspection costs can be reduced by utilizing the predicted contamination level of the hull bottom.

[0080] In addition, according to an embodiment of the present invention, there is an effect of improving the reliability of the predicted hull bottom contamination level through similarity calculation between variable data based on a plurality of factors affecting the hull bottom contamination level.

[0081] The present invention has been described with reference to the embodiments illustrated in the drawings, but this is merely illustrative, and those skilled in the art will understand that various modifications and equivalent alternative embodiments are possible therefrom. Accordingly, the true technical scope of protection of the present invention should be determined by the technical spirit of the appended claims. Explanation of the symbols

[0082] 100: Hull bottom contamination prediction device 110: Input section 120: Data Preprocessing Unit 130: Prediction Unit 140: Learning Section 150: Classification Section 160: Operation section 170: Output section 180: Data Augmentation Unit

Claims

Claim 1 A machine learning-based bottom pollution prediction device for bottom management, comprising: an input unit that receives evaluation target variable data for each vessel regarding multiple factors affecting bottom pollution from a user; a data preprocessing unit that standardizes the evaluation target variable data; and a prediction unit that predicts bottom pollution by applying the standardized evaluation target variable data to a pre-trained learning model. Claim 2 A bottom contamination prediction device according to claim 1, wherein the factors affecting the bottom contamination level include at least one of the following: the start time of bottom contamination, the date of bottom inspection, the number of bottom inspections, the time of mooring of the vessel at the quay, the presence or absence of a typhoon during the mooring period, the mooring period, the mooring quay, the specific gravity of seawater, the seawater temperature, whether cleaning was performed prior to the bottom inspection, the manufacturer of the bottom plating paint, the type of antifouling paint on the vessel's plating, and the antifouling performance of each paint. Claim 3 In claim 1, the data preprocessing unit calculates a Pearson correlation coefficient for each input variable data subject to evaluation, calculates a variance expansion index to remove factors with high correlation between factors affecting the bottom pollution degree, and standardizes the plurality of variable data subject to evaluation into a certain unit, thereby forming a bottom pollution degree prediction device. Claim 4 A bottom contamination prediction device according to claim 1, further comprising: a learning unit that sets learning target variable data as input data and sets bottom contamination degree as output data to train a learning model; a classification unit that classifies evaluation target variable data according to the predicted contamination degree; a calculation unit that calculates the similarity between the evaluation target variable data and the learning target variable data; and an output unit that outputs the predicted bottom contamination degree. Claim 5 In paragraph 4, the output unit further outputs at least one of the variable data of the learning target data most similar to the evaluation target data for which prediction has been completed and the actual pollution value, depending on whether a user requests it. Claim 6 In claim 1, the prediction unit is a bottom contamination prediction device that assigns a high weight to mooring period by quay, mooring period by month, season or antifouling performance among the factors affecting the bottom contamination, and assigns a low weight to water temperature or specific gravity. Claim 7 A bottom contamination prediction device according to claim 4, further comprising a data amplification unit that amplifies the distribution of the learning target data so that the distribution of the classified learning target data is similar to a normal distribution. Claim 8 A bottom contamination prediction device according to paragraph 2, wherein the antifouling performance by paint is classified by comparing the contamination levels between the bottom of the hull and the propeller, and if the contamination level between the bottom of the hull and the propeller differs by 3 levels or more, or if the contamination level of the bottom of the hull is less than 0.5 regardless of the contamination level of the propeller, the antifouling performance by paint is classified as "High"; if the contamination level between the bottom of the hull and the propeller differs by 2 levels or more, or if the contamination level of the bottom of the hull is greater than 0.5 and less than 1 regardless of the contamination level of the propeller, the antifouling performance by paint is classified as "Medium-High"; if the contamination level between the bottom of the hull and the propeller differs by 1 level or more, the antifouling performance by paint is classified as "Medium-Low"; if the contamination level between the bottom of the hull and the propeller is the same, it is classified as "Low"; and if the contamination level of the bottom of the hull is higher than the contamination level of the propeller, the antifouling performance by paint is classified as "Lowest". Claim 9 A method for predicting bottom pollution using a machine learning-based bottom pollution prediction device, comprising the steps of: receiving evaluation target variable data for each vessel regarding multiple factors affecting bottom pollution from a user; standardizing the evaluation target variable data; and applying the standardized evaluation target variable data to a pre-trained learning model to predict bottom pollution. Claim 10 A method for predicting bottom contamination according to claim 9, wherein the factors affecting the bottom contamination include at least one of the following: the start time of bottom contamination, the date of bottom inspection, the number of bottom inspections, the time of mooring of the vessel at the quay, the presence or absence of a typhoon during the mooring period, the mooring period, the mooring quay, the specific gravity of seawater, the seawater temperature, whether cleaning was performed prior to the bottom inspection, the manufacturer of the bottom plating paint, the type of antifouling paint on the vessel's plating, and the antifouling performance of each paint. Claim 11 In claim 9, the step of standardizing the evaluation target variable data comprises calculating a Pearson correlation coefficient for each input evaluation target variable data, calculating a variance expansion index to remove factors with high correlation between factors affecting the hull bottom contamination degree, and standardizing the plurality of evaluation target variable data into a certain unit, thereby predicting the hull bottom contamination degree. Claim 12 A method for predicting bottom contamination according to claim 9, further comprising the steps of: setting learning target variable data as input data and setting bottom contamination degree as output data to train a learning model; classifying evaluation target variable data according to the predicted contamination degree; calculating similarity between the evaluation target variable data and the learning target variable data; and outputting the predicted bottom contamination degree. Claim 13 In claim 12, the step of outputting the predicted bottom contamination level is a bottom contamination level prediction method that outputs at least one of the data of the learning target data most similar to the evaluation target data for which the prediction is completed and the actual contamination level value, depending on whether a user requests it. Claim 14 In claim 9, the step of predicting bottom contamination by applying the standardized evaluation target variable data to a pre-trained learning model is a bottom contamination prediction method in which high weights are assigned to mooring periods by quay, monthly mooring periods, seasons, or antifouling performance among the factors affecting bottom contamination, and low weights are assigned to water temperature or specific gravity. Claim 15 A method for predicting bottom contamination according to claim 12, further comprising the step of amplifying the distribution of the learning target data so that the distribution of the classified learning target data is similar to a normal distribution. Claim 16 A method for predicting bottom contamination levels according to claim 10, wherein the antifouling performance of each paint is classified by comparing the contamination levels between the bottom of the hull and the propeller, and if the contamination level between the bottom of the hull and the propeller differs by 3 levels or more, or if the contamination level of the bottom of the hull is less than 0.5 regardless of the contamination level of the propeller, the antifouling performance of each paint is classified as "High"; if the contamination level between the bottom of the hull and the propeller differs by 2 levels or more, or if the contamination level of the bottom of the hull is greater than 0.5 and less than 1 regardless of the contamination level of the propeller, the antifouling performance of each paint is classified as "Medium-High"; if the contamination level between the bottom of the hull and the propeller differs by 1 level or more, the antifouling performance of each paint is classified as "Medium-Low"; if the contamination level between the bottom of the hull and the propeller is the same, it is classified as "Low"; and if the contamination level of the bottom of the hull is higher than the contamination level of the propeller, the antifouling performance of each paint is classified as "Lowest".

Images