System for removing green algae pixel in the candidate group of sargassum horneri pixel through analysis of color-geostationary satellite images and method thereof

The system addresses false detections in Sargassum monitoring by employing geostationary satellite image analysis with preprocessing and determination units to accurately classify Sargassum pixels, improving precision and objectivity in satellite image analysis.

KR103046774B1Active Publication Date: 2026-11-11UNDERWATER SURVEY TECH
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
KR1020240090300
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Filing Date
2024-07-09
Publication Date
2026-11-11
Estimated Expiration
2044-04-04

AI Technical Summary

Technical Problem

Existing methods for monitoring Sargassum horneri using satellite imagery suffer from false detections, which are time-consuming and subjective, leading to inaccurate identification and misidentification of green algae and cloud areas.

Method used

A system and method for removing green algae pixels through geostationary satellite image analysis, utilizing preprocessing, first and second determination units, and a false detection removal unit, including modules for cloud area detection, dominant image determination, and band-specific reflectance value comparisons to accurately classify Sargassum horneri pixels.

Benefits of technology

Provides accurate monitoring results by automatically removing false detections, distinguishing between Sargassum and green algae, and preventing cloud misidentification, thereby enhancing precision and objectivity in satellite image analysis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 112024074328824-PAT00015_ABST
    Figure 112024074328824-PAT00015_ABST
Patent Text Reader

Abstract

The present invention relates to a system and method for removing green algae pixels within a candidate group of Sargassum horneri through analysis of geostationary satellite imagery. More specifically, the invention relates to a system and method for removing green algae pixels within a candidate group of Sargassum horneri through analysis of geostationary satellite imagery that provides accurate monitoring results to a user by removing the green algae detection area when monitoring Sargassum horneri through satellite imagery.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present invention relates to a system and method for removing green algae pixels within a candidate group of Sargassum horneri through analysis of geostationary satellite imagery. More specifically, the invention relates to a system and method for removing green algae pixels within a candidate group of Sargassum horneri through analysis of geostationary satellite imagery that provides accurate monitoring results to a user by removing the green algae detection area when monitoring Sargassum horneri through satellite imagery. Background Technology

[0003] Sargassum horneri, a brown algae that grows on intertidal reefs, is known to have numerous air sacs morphologically, allowing it to float on the surface of the seawater and travel long distances even after detaching from the rocks.

[0004] For the past few years, large quantities of alien Sargassum horneri, which grow between winter and early summer, have been flowing into and accumulating along the domestic coast. As this accumulated Sargassum decomposes, it causes foul odors and attracts various pests. Furthermore, it damages the tourism industry by spoiling the landscape and creates various problems, such as hindering the operation of small vessels. Additionally, it can lead to various issues, including incurring additional costs for processing the Sargassum that has flowed into the domestic coast.

[0005] To handle the inflow of Sargassum, a method of continuously monitoring or detecting the Sargassum is required, and satellite image analysis is known to be the most effective means. However, various types of false detections can occur when monitoring Sargassum through satellite image analysis. Until now, a method has been used in which workers visually remove false detections of Sargassum; however, visual interpretation requires a significant amount of time to eliminate false detections, and since it involves the subjectivity of the worker, differences in precision are inevitable depending on their skill level.

[0007] In order to solve such problems, the inventors of the present invention intend to present a system and method for removing green algae pixels within a candidate group of Sargassum horneri through the analysis of a novel geostationary satellite imagery with an improved method, and details thereof will be described below. Prior art literature

[0009] Korean Registered Patent No. 10-1737354 'Method and System for Detecting Seaweed on the Seawater Surface Using Optical Satellite Imagery' The problem to be solved

[0010] It was devised to solve the problems of the prior art described above,

[0012] The present invention aims to provide a system and method for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis, which provides accurate results to the user by removing previously classified false detection types when monitoring Sargassum through satellite imagery.

[0013] In addition, the present invention aims to provide a system and method for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis, which enables accurate identification of floating algae pixels within a satellite image through a first determination unit.

[0014] In addition, the present invention aims to provide a system and method for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis, which accurately removes cloud areas within a satellite image through a second judgment unit and thereby prevents the outer parts of clouds or thin clouds from being misidentified as Sargassum.

[0015] In addition, the present invention aims to provide a system and method for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis, which automatically removes type-specific false detection pixels among the Sargassum candidate group pixels through a false detection removal unit.

[0016] In addition, the present invention aims to provide a system and method for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis, which enables accurate false detections to be removed according to the appearance rate of Sargassum by period through a sporadic false detection removal module that sets different false detection removal conditions by period.

[0017] In addition, the present invention aims to provide a system and method for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis, which enables objective derivation of false detections by determining the starting point for changing conditions for sporadic false detections within the initial occurrence period through spatial density rather than user subjectivity.

[0018] In addition, the present invention aims to provide a system and method for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis, which allows for more precise removal of green algae detection results by providing a green algae detection removal module that sets the method of classifying Sargassum / green algae differently depending on whether the image is Sargassum dominant or green algae dominant. means of solving the problem

[0020] The present invention may be implemented by an embodiment having the following configuration to achieve the aforementioned objectives.

[0022] According to one embodiment of the present invention, a system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to the present invention comprises: a preprocessing unit for preprocessing a satellite image; a first determination unit for determining floating algae pixels by considering pixel-specific reflectance values ​​in the preprocessed satellite image; a second determination unit for determining the remaining pixels as Sargassum candidate group pixels, excluding pixels determined as cloud areas among the floating algae pixels identified by the first determination unit through cloud area detection; and a false detection removal unit for automatically removing type-specific false detection pixels among the Sargassum candidate group pixels; wherein the false detection removal unit comprises a green algae detection removal module for removing green algae detection areas within the satellite image in which Sargassum candidate group pixels are identified through the second determination unit.

[0023] According to another embodiment of the present invention, the algal bloom detection and removal module in the system for removing algal bloom pixels within a Sargassum candidate group through geostationary satellite image analysis according to the present invention is characterized by including a dominant image determination module that determines whether the image is Sargassum dominant or algal bloom dominant by comparing the magnitudes of the reflectance values ​​of the green band wavelength and the red band wavelength for each pixel of the Sargassum candidate group pixels within the satellite image.

[0024] According to another embodiment of the present invention, the algal bloom detection and removal module in the system for removing algal bloom pixels within a Sargassum candidate group through geostationary satellite image analysis according to the present invention further comprises a band-specific reflectance value comparison module that compares the reflectance values ​​of the green band wavelength and the red band wavelength of individual pixels identified as a Sargassum candidate group within the satellite image determined as the Sargassum dominant image and the algal bloom dominant image.

[0025] According to another embodiment of the present invention, the band-specific reflectance value comparison module in the system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to the present invention is characterized by determining a pixel as a Sargassum pixel when the reflectance value of the red band wavelength at a pixel identified as a Sargassum candidate group within a Sargassum dominant image is greater than the reflectance value of the green band wavelength.

[0026] According to another embodiment of the present invention, the algae detection and removal module in the system for removing algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to the present invention further comprises a re-determination module that re-determines the suitability of a determined pixel when it is determined to be an algae pixel in an image dominant in Sargassum, or when it is determined to be a Sargassum pixel in an image dominant in algae.

[0027] According to another embodiment of the present invention, the recrystallization module in the system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to the present invention is characterized by determining, when a pixel is determined to be a green algae pixel within a Sargassum dominant image, that the reflectance value of the green band wavelength minus the reflectance value of the red band wavelength is greater than or equal to a second threshold value, and that the pixel is determined to be a Sargassum pixel if it is less than the second threshold value.

[0028] According to another embodiment of the present invention, the recrystallization module in the system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to the present invention is characterized by determining, when a Sargassum pixel is determined within an image of green algae dominance, that the reflectance value of the red band wavelength minus the reflectance value of the green band wavelength is less than or equal to a third threshold value, and confirming it as a Sargassum pixel when it exceeds the third threshold value. Effects of the invention

[0030] The present invention has the following effects based on the configuration described above.

[0032] The present invention has the effect of providing accurate results to the user by removing previously classified false detection types when monitoring Sargassum horneri through satellite imagery.

[0033] In addition, the present invention has the effect of enabling accurate identification of floating algae pixels within a satellite image through a first determination unit.

[0034] In addition, the present invention provides the effect of accurately removing cloud areas within a satellite image through a second judgment unit, thereby preventing the outer parts of clouds or thin clouds from being misidentified as Sargassum.

[0035] In addition, the present invention has the effect of automatically removing type-specific false detection pixels among the Sargassum candidate pixels through a false detection removal unit.

[0036] In addition, the present invention demonstrates the effect of enabling accurate false detections to be eliminated according to the appearance rate of Sargassum horneri through a sporadic false detection removal module that sets different false detection removal conditions for each period.

[0037] In addition, the present invention has the effect of enabling objective derivation of false positives by determining the starting point for changing the application conditions for sporadic false positives within the initial occurrence period through spatial density rather than user subjectivity.

[0038] In addition, the present invention has the effect of removing algae detection results more precisely by providing an algae detection removal module that sets the method of classifying Sargassum / algae differently depending on whether the image is Sargassum dominant or algae dominant. Brief explanation of the drawing

[0040] FIG. 1 is a conceptual diagram of a system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to an embodiment of the present invention; FIG. 2 is a block diagram of a system for removing green algae pixels within a Sargassum candidate group through analysis of geostationary satellite imagery according to FIG. 1; FIG. 3 is a block diagram of a preprocessing unit according to FIG. 2; FIG. 4 is a block diagram of a first judgment unit according to FIG. 2; FIG. 5 is a block diagram of a second judgment unit according to FIG. 2; FIG. 6 is a reference diagram for explaining a cloud area detection module according to FIG. 5; FIG. 7 is a block diagram of a false detection removal unit according to FIG. 2; FIG. 8 is a reference diagram for explaining the false detection type according to FIG. 7; FIG. 9 is a block diagram of a sporadic false positive removal module according to FIG. 7; FIG. 10 is a block diagram of an algae removal module according to FIG. 7; FIG. 11 is a flowchart of a method for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to an embodiment of the present invention; FIG. 12 is a flowchart for the preprocessing step according to FIG. 11; FIG. 13 is a flowchart for the first determination step according to FIG. 11; FIG. 14 is a flowchart for the second determination step according to FIG. 11; FIG. 15 is a flowchart for the false positive removal step according to FIG. 11; FIG. 16 is a flowchart for the sporadic false positive removal step according to FIG. 15; Figure 17 is a flowchart of the algae detection and removal step according to Figure 15. Specific details for implementing the invention

[0041] Hereinafter, embodiments of the present invention will be described in more detail with reference to the accompanying drawings. Embodiments of the present invention may be modified in various forms, and the scope of the present invention should not be interpreted as being limited to the embodiments below, but should be interpreted based on the matters described in the claims. Furthermore, these embodiments are provided merely for reference to more completely explain the present invention to those with average knowledge in the art.

[0042] As used herein, the singular form may include the plural form unless the context clearly indicates otherwise. Additionally, as used herein, “comprise” and / or “comprising” specify the presence of the mentioned features, numbers, steps, actions, parts, elements, and / or groups thereof, and do not exclude the presence or addition of one or more other features, numbers, actions, parts, elements, and / or groups.

[0043] In this specification, it should be noted that individual components may be formed integrally or independently as needed, and there are no separate limitations thereon.

[0045] FIG. 1 is a conceptual diagram of a system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to one embodiment of the present invention; FIG. 2 is a block diagram of a system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to FIG. 1.

[0047] Hereinafter, a system (1) for removing green algae pixels within a candidate group of Sargassum horneri through analysis of geostationary satellite imagery according to one embodiment of the present invention will be described in detail with reference to the attached drawings.

[0049] Referring to FIGS. 1 and 2, the present invention relates to a system (1) for removing green algae pixels within a candidate group of Sargassum horneri through analysis of geostationary satellite imagery, and more specifically, to a system (1) for removing green algae pixels within a candidate group of Sargassum horneri through analysis of geostationary satellite imagery that provides accurate monitoring results to a user by removing the green algae detection area when monitoring Sargassum horneri through satellite imagery.

[0050] To this end, the system (1) for removing green algae pixels within a candidate group of Sargassum horneri through the analysis of the aforementioned geostationary satellite imagery may include a preprocessing unit (10), a first judgment unit (30), a second judgment unit (50), and a false detection removal unit (70).

[0052] Figure 3 is a block diagram of the preprocessing unit according to Figure 2.

[0054] Referring to FIGS. 1 to 3, the preprocessing unit (10) is configured to preprocess satellite imagery for monitoring Sargassum horneri. The system (1) for removing green algae pixels within a candidate group of Sargassum horneri through geostationary ocean color satellite image analysis according to one embodiment of the present invention may utilize images captured by, for example, GOCI-II (Geostationary Ocean Color Imager-II) mounted on the Cheollian 2B satellite, but the scope of the present invention is not limited thereto.

[0055] And the preprocessing unit (10) may include a reflectance conversion module (110), a molecular scattering correction module (130), and a surrounding area removal module (150).

[0056] The reflectance conversion module (110) is a module that converts the Top Of Atmosphere (TOA) radiance value into a reflectance value for the top of the atmosphere. In one embodiment of the present invention, since Sargassum horneri is identified in a satellite image based on the reflectance value, a configuration that converts the radiance value into a reflectance value by the reflectance conversion module (110) is essential.

[0057] This reflectance conversion module (110) is, for example,

[0059] (1) a(λ) = (π * b(λ)) / (c(λ) * cos(d))

[0060] (2) c(λ) = e(λ) * f(g)

[0062] Through this, the radiation value of the upper atmosphere can be converted into a reflectance value. In Equation (1), λ is a specific wavelength, a(λ) is the reflectance of the upper atmosphere, b(λ) is the radiation of the upper atmosphere, c(λ) is the solar radiation incident on Earth, and d is the solar zenith angle; in Equation (2), e is the solar radiation in space without considering the Sun-Earth distance, f is the Sun-Earth distance coefficient, and g is a value representing the day of the year.

[0063] The molecular scattering correction module (130) is a module that removes the reflectance (Rayleigh Reflectance) caused by atmospheric molecular scattering from the reflectance value of the upper atmosphere converted through the reflectance conversion module (110). Generally, the reflectance value of the upper atmosphere, which is the value of a (λ) converted through the reflectance conversion module (110), includes atmospheric influences, and in order to identify Sargassum horneri floating on the surface of the seawater, the atmospheric influence must be removed. Therefore, the molecular scattering correction module (130) removes the reflectance caused by atmospheric molecular scattering, and the specific process is as shown in Equation (3) below, for example.

[0065] (3) h(λ) = a(λ) - i(λ)

[0067] Equation (3) is the reflectance value with atmospheric effects removed, and i(λ) is the reflectance value due to atmospheric molecular scattering.

[0068] The surrounding area removal module (150) is a module that calculates and removes surrounding areas unrelated to Sargassum detection in a satellite image. For example, the surrounding area removal module (150) can specifically exclude land areas, clouds, and, if necessary, turbid seawater areas that are unrelated to Sargassum detection within the satellite image. This surrounding area removal can be performed using wavelength-specific reflectance values ​​at individual pixels of the satellite image, and there are no separate limitations thereon.

[0070] FIG. 4 is a block diagram of the first judgment unit according to FIG. 2.

[0072] Referring to FIGS. 1, 2 and 4, the first determination unit (30) is configured to determine floating algae pixels by utilizing pixel-by-pixel reflectance values ​​within a preprocessed satellite image. To this end, the first determination unit (30) may include a first determination module (310), a second determination module (330), and an occupancy area ratio calculation module (350).

[0073] The first determination module (310) is a module that determines floating algae candidate pixels by calculating the difference in reflectance values ​​for each pixel and wavelength that make up the preprocessed satellite image. The first determination module (310) can determine floating algae candidate pixels, for example, through the following equation (4).

[0075] (4) RE Diff = R rc,NIR - R rc,RED

[0077] R in Equation (4) rc,NIR is the Rayleigh-corrected reflectance value in the near-infrared wavelength band per pixel, and R rc,RED is the Rayleigh-corrected reflectance value in the red wavelength band for each pixel. Since floating algae pixels have high reflectance values ​​in the near-infrared band, equation (4) is calculated for each pixel, RE Diff If the value is greater than or equal to the set value, it can be determined as a floating algae candidate pixel.

[0078] The second determination module (330) is a module that identifies floating algae pixels by calculating the difference in reflectance between the target pixel identified as a floating algae candidate pixel by the first determination module (310) and the average reflectance of the background seawater pixel. This second determination module (330) can identify floating algae pixels by calculating the difference in reflectance of the target pixel and the background seawater pixel at a preset band wavelength. Through the second determination module (330), floating algae pixels among the target pixels identified as floating algae candidate pixels by the first determination module (310) can be identified.

[0079] The occupancy area ratio calculation module (350) is a module that calculates the occupancy area ratio of floating algae within individual pixels identified as floating algae pixels by the second determination module (330).

[0082] FIG. 5 is a block diagram of the second judgment unit according to FIG. 2; FIG. 6 is a reference diagram for explaining the cloud area detection module according to FIG. 5.

[0084] Referring to FIGS. 1, 2 and 5, the second determination unit (50) is configured to determine whether floating algae pixels confirmed through the first determination unit (30) are Sargassum candidate group pixels. To this end, the second determination unit (50) may include a cloud area detection module (510) and a Sargassum candidate group determination module (530).

[0085] Referring to FIGS. 1 and 2 and FIGS. 5 and 6, the cloud area detection module (510) is a module that merges cloud data (or data) from a plurality (preferably two) of satellite image data and additionally removes cloud areas within a satellite image preprocessed by the preprocessing unit (10). As described above, the preprocessed satellite image may be one taken by GOCI-II mounted on the Cheollian 2B satellite. At this time, the cloud area is removed by the surrounding area removal module (150) during the preprocessing process of the preprocessed satellite image, but cases may occur where the cloud area is not completely removed because the outer part of the cloud or thin clouds are not detected. If the cloud area is not completely removed in this way, the outer part of the cloud or thin clouds may be misidentified as Sargassum horneri. In particular, since thin clouds exhibit spectral characteristics similar to Sargassum horneri floating on the sea surface, there is a high possibility of misidentification.

[0086] In order to prevent such problems, in one embodiment of the present invention, cloud regions within a preprocessed satellite image are additionally removed through a cloud region detection module (510). At this time, in order to additionally remove cloud regions within the preprocessed satellite image, cloud detection data from heterogeneous satellite images may be utilized. For example, cloud detection data from Level 2 data of the Chollian 2A Advanced Meteorological Imager (AMI) may be utilized. Hereinafter, the above 'heterogeneous satellite image' is referred to as the 'target image'.

[0087] To analyze cloud regions for preprocessed satellite images by merging cloud data from target images, conditions for determining cloud regions can be established. For example, pixel-by-pixel cloud regions (Cloud) or ocean regions (Sea) within preprocessed satellite images such as GOCI-II can be classified, and pixel-by-pixel regions likely to be cloud (Probably Cloud), cloud regions (Cloud), and ocean regions (Sea) within the target image can be classified. In this case, each region can be classified based on the reflectance value of a pre-set band (e.g., infrared band). Through this, six conditions can be derived from multiple satellite images (Sea-Sea, Sea-Probably Cloud, Sea-Cloud, Cloud-Sea, Cloud-Probably Cloud, Cloud-Cloud). Additionally, flags can be set for each of the six conditions. Furthermore, pixel-by-pixel flags within the preprocessed satellite images can be assigned by utilizing cloud detection data from the target images. Individual pixels within the preprocessed satellite images can be determined to be cloud regions, with the exception of Sea-Sea among the aforementioned six conditions. Figure 6a is a preprocessed satellite image, Figure 6b is a target image, Figure 6c is a preprocessed satellite image with cloud data of the target image merged, and Figure 6d is a table to explain the flag setting criteria for six conditions.

[0088] The Sargassum candidate group determination module (530) is a module that determines the remaining pixels as Sargassum candidate group pixels, excluding the pixels determined as cloud areas by the cloud area detection module (510) among the pixels determined as floating algae pixels through the first determination unit (30). Therefore, the Sargassum pixel is not determined through the Sargassum candidate group determination module (530), but rather the Sargassum candidate group pixels that are likely to be Sargassum pixels can be determined.

[0090] FIG. 7 is a block diagram of a false detection removal unit according to FIG. 2; FIG. 8 is a reference diagram for explaining the types of false detection according to FIG. 7; FIG. 9 is a block diagram of a sporadic false detection removal module according to FIG. 7; and FIG. 10 is a block diagram of an algae removal module according to FIG. 7.

[0092] Referring to FIGS. 1, 2 and FIGS. 7 to 8, the false detection removal unit (70) is configured to automatically remove type-specific false detection pixels among the Sargassum candidate group pixels confirmed through the second judgment unit (50). The types of false detection considered by the false detection removal unit (70) may be one or more of linear false detection (see FIG. 8a), sporadic false detection (see FIG. 8b), and green algae detection (see FIG. 8c) pixels. To this end, the false detection removal unit (70) may include a linear false detection removal module (710), a sporadic false detection removal module (730), and a green algae detection removal module (750).

[0093] The linear false detection removal module (710) is a module that removes linear false detection areas within a satellite image in which Sargassum candidate pixels have been confirmed through the second judgment unit (50). 'Linear false detection' occurs in the form of a line in the Sargassum detection result and has spectral characteristics similar to Sargassum, and may occur in areas where the Sargassum is not mainly distributed. In addition, since Sargassum has a high density, the linear false detection removal module (710) can perform masking by determining that a linear area among the Sargassum candidate pixels is not a Sargassum pixel.

[0094] Referring to FIGS. 1 to 2 and FIGS. 7 to 9, the sporadic false detection removal module (730) is a module that removes sporadic false detection areas within a satellite image in which Sargassum candidate pixels are confirmed through the second judgment unit (50). 'Sporadic false detection' mainly occurs in moving objects such as thin clouds or ships due to differences in observation time by wavelength band, or in objects existing above the sea surface such as aquaculture farms or offshore wind turbines that have spectral characteristics similar to Sargassum. Therefore, in order to accurately monitor Sargassum, sporadic false detection areas must be removed, and to this end, the sporadic false detection removal module (730) may include a period setting module (731), a period window setting module (733), a density calculation module (735), and a false detection area removal module (737).

[0095] The period setting module (731) is a module that classifies periods based on the occurrence rate of Sargassum horneri in order to precisely remove sporadic false detection areas among Sargassum horneri candidate pixels. Generally, floating Sargassum horneri rarely appears from September to November, occurs initially from December to February, and occurs in large quantities from March to August. Therefore, the period setting module (731) can classify periods based on the occurrence rate of Sargassum horneri into a 'period of non-occurrence', an 'period of initial occurrence (or period of possible occurrence)', and a 'period of large occurrence'.

[0096] The period-specific window setting module (733) is a module that sets the period-specific window size classified through the period setting module (731). For example, a window of size 3 * 3 can be set for the non-appearance period, a window of size 3 * 3 can be set for the initial occurrence period, and a window of size 5 * 5 can be set for the mass appearance period, but the scope of the present invention is not limited thereto.

[0097] The density calculation module (735) is a module that analyzes the spatial density of Sargassum candidate groups to automatically determine the point in time to change the sporadic false detection area removal condition applied during the initial occurrence period from Condition 2 to Condition 3 below. Therefore, even if the initial occurrence period and the mass appearance period are set through the period setting module (731), the point in time to change from Condition 2 to Condition 3 can be determined through the density calculation module (735). To this end, the density calculation module (735) can analyze the spatial density of Sargassum candidate group pixels within the satellite image through a kernel density estimation method. If such spatial density is greater than or equal to a pre-set first threshold value (Th1), Condition 2 below is changed to Condition 3, and after changing to Condition 3, it is desirable to continue applying Condition 3 without recalculating the spatial density until the non-appearance period has passed.

[0098] The false detection area removal module (737) is a module that removes sporadic false detection areas among Sargassum candidate pixels through the period classified through the period setting module (731) and the period-specific window size set through the period-specific window setting module (733). For example, during the non-occurrence period, if there are n1 or fewer Sargassum candidate pixels within a 3 * 3 window, the said pixels are removed from the result (hereinafter referred to as 'Condition 1'). Also, during the initial occurrence period, if there are n2 or fewer Sargassum candidate pixels within a 3 * 3 window, the said pixels are removed from the result (hereinafter referred to as 'Condition 2'). Finally, during the mass occurrence period, if there are n3 or fewer Sargassum candidate pixels within a 5 * 5 window, the said pixels are removed from the result (hereinafter referred to as 'Condition 3'). As described above, the point in time when changing from condition 2 to condition 3 can be determined by the density calculation module (735).

[0099] n1, n2, and n3 are all natural numbers, for example, n1=2, n2=1, and n3=1. In order to precisely eliminate sporadic false detections based on the occurrence rate of Sargassum horneri, it is desirable that the window size of the initial occurrence period is larger than the window size of the non-occurrence period, or that n2 < n1 is satisfied when the window size is the same. Additionally, it is desirable that the window size of the mass occurrence period is larger than the initial occurrence period, or that n3 < n2 is satisfied. Alternatively, the false detection area removal module (737) may set a standard so that the initial occurrence period is not judged as a false detection area even when there is a higher density of Sargassum horneri within the set window size compared to the non-occurrence period, and may set a standard so that the mass occurrence period is not judged as a false detection area even when there is a higher density of Sargassum horneri within the set window size compared to the initial occurrence period.

[0100] Referring to FIGS. 1 to 2, FIGS. 7 to 8 and FIG. 10, the algal bloom detection removal module (750) is a module that removes the algal bloom detection area from a satellite image in which the Sargassum candidate group pixels are determined through the second determination unit (50). To this end, the algal bloom detection removal module (750) may include a dominant image determination module (751), a band-specific reflectance value comparison module (753), and a recrystallization module (755).

[0101] The dominant image determination module (751) is a module that determines whether the image is a Sargassum dominant image or an algal bloom dominant image by comparing the reflectance values ​​of the green wavelength band and the red wavelength band for each pixel among the Sargassum candidate pixels in the satellite image. Generally, algal blooms show high reflectance values ​​in the green wavelength band of 550 nm, and Sargassum has high reflectance values ​​in the red wavelength band between 620 and 660 nm.

[0102] Considering this, by comparing the reflectance values ​​of the green and red wavelength bands of individual pixels forming the Sargassum candidate pixels within the satellite image, if the number of pixels having a higher reflectance value in the green wavelength band than in the red wavelength band is greater than in the opposite case, the satellite image in which the Sargassum candidate pixels were captured can be determined as an 'algae-dominant image'. Conversely, if the number of pixels having a higher reflectance value in the red wavelength band than in the green wavelength band is greater than in the opposite case, the satellite image in which the Sargassum candidate pixels were captured can be determined as an 'algae-dominant image'.

[0103] The band-specific reflectance value comparison module (753) is a module that compares the magnitude of the reflectance values ​​of the green band and the red band of individual pixels identified as Sargassum candidates within a satellite image determined to be a Sargassum dominant image and an algal bloom dominant image. For example, if the reflectance value of the red wavelength band in an individual pixel identified as a Sargassum candidate within a Sargassum dominant image is higher than the reflectance value of the green wavelength band, the pixel may be determined to be a Sargassum pixel. Conversely, if the reflectance value of the red wavelength band in a pixel identified as a Sargassum candidate within a Sargassum dominant image is smaller than the reflectance value of the green wavelength band (i.e., determined to be an algal bloom pixel), the recrystallization module (755) may be operated.

[0104] And if the reflectance value of the green wavelength band at a pixel identified as a candidate for Sargassum in the green algae dominant image is greater than the reflectance value of the red wavelength band, it can be determined as a green algae pixel, and if the opposite is true (i.e., determined as a Sargassum pixel), the recrystallization module (755) can be operated.

[0105] The recrystallization module (755) is a module that recrystallizes the suitability of a determined pixel when it is determined to be an algae pixel in an image dominated by Sargassum, or conversely, when it is determined to be a Sargassum pixel in an image dominated by algae. To this end, the recrystallization module (755) determines whether it is a Sargassum pixel or an algae pixel by comparing the reflectance value in the red wavelength band and the difference value in the reflectance in the green wavelength band of the individual pixel to be applied.

[0106] For example, if the recrystallization module (755) determines that a pixel is green in a Sargassum dominant image, it is determined to be a green pixel if the reflectance value of the green wavelength band minus the reflectance value of the red wavelength band is greater than or equal to the second threshold value (Th2), and a Sargassum pixel if it is less than the second threshold value (Th2). Additionally, if the recrystallization module (755) determines that a pixel is green in a Sargassum dominant image, it is determined to be a green pixel if the reflectance value of the red wavelength band minus the reflectance value of the green wavelength band is less than or equal to the third threshold value (Th3), and a Sargassum pixel if it is greater than the third threshold value (Th3). The second threshold value (Th2) and the third threshold value (Th3) are values ​​that can be determined experimentally, for example, the second threshold value (Th2) may be 0.00038 and the third threshold value (Th3) may be 0.00058. In this way, to determine whether there are Sargassum pixels or green algae pixels, double or, in some cases, triple conditions are applied, making it possible to remove green algae detection pixels more accurately.

[0108] FIG. 11 is a flowchart of a method for monitoring Sargassum horneri through satellite image analysis according to an embodiment of the present invention.

[0110] Hereinafter, a method (S1) for removing green algae pixels within a candidate group of Sargassum horneri through analysis of geostationary satellite imagery according to an embodiment of the present invention will be described in detail with reference to the attached drawings. Since this method is performed through the aforementioned system (1), detailed information regarding each step / configuration will be omitted.

[0112] Referring to FIG. 11, a method (S1) for removing green algae pixels within a candidate group of Sargassum horneri through geostationary satellite image analysis according to one embodiment of the present invention may include a preprocessing step (S10), a first determination step (S30), a second determination step (S50), and a false detection removal step (S70).

[0114] Figure 12 is a flowchart of the preprocessing step according to Figure 11.

[0116] Referring to FIG. 12, the preprocessing step (S10) is a step of preprocessing a satellite image. To explain the preprocessing step (S10) in detail, first, a step of converting the upper atmosphere (TOA) radiation value into a reflectance value for the upper atmosphere (S110), a step of removing the reflectance value due to atmospheric molecular scattering from the reflectance value of the upper atmosphere converted through step S110 (S130), and a step of removing surrounding areas unrelated to Sargassum detection (S150) may be performed.

[0118] FIG. 13 is a flowchart for the first determination step according to FIG. 11.

[0120] Referring to FIG. 13, when the preprocessing step (S10) is completed, the first determination step (S30) can be performed. The first determination step (S30) is a step of determining floating algae pixels using the preprocessed satellite image. The first determination step (S30) can be performed as follows.

[0121] First, a step (S310) may be performed to determine floating algae candidate pixels by calculating the difference in reflectance values ​​for each pixel and wavelength constituting the preprocessed satellite image. Step S310 may determine floating algae candidate pixels by, for example, calculating the difference in reflectance values ​​between the Rayleigh-corrected near-infrared wavelength band and the Rayleigh-corrected red wavelength band. Then, a step (S330) may be performed to identify floating algae pixels by calculating the difference in average reflectance between the target pixels identified as floating algae candidate pixels and the background seawater pixels. Subsequently, a step (S350) may be performed to calculate the ratio of the occupied area of ​​floating algae within individual pixels among the identified floating algae pixels.

[0123] FIG. 14 is a flowchart for the second judgment step according to FIG. 11.

[0125] Referring to FIG. 14, when the first determination step (S30) is completed, the second determination step (S50) may be performed. The second determination step (S50) is a step for determining whether the floating algae pixels confirmed through the first determination step (S30) are Sargassum candidate pixels. The second determination step (S50) may proceed as follows. First, cloud areas within the preprocessed satellite image can be additionally removed using cloud data of the target image (S510). After step S510, among the pixels determined as floating algae pixels, pixels excluding those determined as cloud areas by step S510 can be determined as Sargassum candidate pixels (S530).

[0127] Figure 15 is a flowchart of the false detection removal step according to Figure 11.

[0129] Referring to FIG. 15, when a candidate pixel of Sargassum horneri is determined, a false detection removal step (S70) for removing false detection areas may be performed. As previously mentioned, the types of false detection may include one or more of linear false detections, sporadic false detections, and algal bloom detections. To this end, it should be noted that the false detection removal step (S70) may include a linear false detection removal step, a sporadic false detection removal step, and an algal bloom detection step as described below, and that these three steps may proceed in parallel with each other.

[0130] First, linear false positive regions within the satellite image in which Sargassum candidate pixels are identified are removed (S710). That is, through step S710, regions occurring in the form of lines among the Sargassum candidate pixels within the satellite image can be removed.

[0132] Figure 16 is a flowchart of the sporadic false positive removal step according to Figure 15.

[0134] Referring to FIG. 16, sporadic false detection areas within the satellite image in which Sargassum candidate pixels are identified are removed (S730). Step S730 is described in detail as follows: the date the satellite image being analyzed was captured is determined to be a point in time among the non-appearance period, the initial occurrence period, and the mass occurrence period (S731); and for satellite images captured during the initial occurrence period, the spatial density of Sargassum candidate pixels is analyzed (S733). Step S733 can be performed using a kernel density estimation method. Subsequently, whether the spatial density obtained through Step S733 is greater than or equal to a first threshold value (Th1) is determined (S735).

[0135] Through steps S731 to S735, it can be determined which of conditions 1 to 3 is applied to the satellite image being analyzed.

[0136] Then, sporadic false detection areas can be removed through conditions (any one of conditions 1 to 3) applied to the satellite image (S737). Since step S737 is performed through the false detection area removal module (737), a detailed description thereof will be omitted.

[0138] Figure 17 is a flowchart of the algae detection and removal step according to Figure 15.

[0140] Additionally, referring to FIG. 17, the green algae detection area within the satellite image in which the Sargassum candidate group pixels are identified is removed (S750). Step S750 is described in detail below. First, the Sargassum candidate group pixels within the satellite image being analyzed are analyzed to determine whether the image is predominantly Sargassum or predominantly green (S751).

[0141] If the image is determined to be predominantly Sargassum by step S751, the reflectance values ​​of the green wavelength band and the red wavelength band for each candidate pixel of Sargassum are compared (S752). If the reflectance value of the green wavelength band is higher by step S752, the process proceeds to step S753, and if the opposite is true, the pixel is confirmed as a Sargassum pixel (S757).

[0142] In addition, in step S753, if the green wavelength band reflectance value of a Sargassum candidate pixel minus the red wavelength band reflectance value is greater than or equal to the second threshold value (Th2), the pixel is confirmed as an algae pixel and removed (S756), and if the opposite is true, the pixel is confirmed as a Sargassum pixel (S757).

[0143] Then, if the image is determined to be predominantly green by step S751, the reflectance values ​​of the green wavelength band and the red wavelength band for each pixel of the Sargassum horneri candidate group are compared (S754). If the reflectance value of the red wavelength band is greater by step S754, the process proceeds to step S755, and if the opposite is true, the pixel is confirmed as a green pixel and removed (S756).

[0144] And in step S755, if the red wavelength band reflectance value of a Sargassum candidate pixel minus the green wavelength band reflectance value is less than or equal to the third threshold value (Th3), the pixel is confirmed as an algae pixel and removed (S756), and if the opposite is true, the pixel is confirmed as a Sargassum pixel (S757).

[0146] The above detailed description is illustrative of the present invention. Furthermore, the foregoing describes preferred embodiments of the present invention, and the present invention may be used in various other combinations, modifications, and environments. That is, modifications or alterations are possible within the scope of the concept of the invention disclosed herein, the scope equivalent to the written disclosure, and / or the scope of the art or knowledge. The foregoing embodiments describe the best state for implementing the technical concept of the present invention, and various modifications required for specific fields of application and uses of the present invention are also possible. Accordingly, the above detailed description of the invention is not intended to limit the present invention to the disclosed embodiments. Explanation of the symbols

[0148] 1: System for Removing Green Algae Pixels from Sargassum Candidates via Geostationary Ocean Color Satellite Image Analysis 10 : Preprocessing section 110: Reflectance conversion module 130: Molecular scattering correction module 150 : Peripheral area removal module 30: First Judgment Division 310: 1st judgment module 330: 2nd judgment module 350 : Occupancy Area Ratio Calculation Module 50 : 2nd Judgment Division 510: Cloud area detection module 530: Sargassum candidate determination module 70 : False positive removal unit 710: Linear False Positive Elimination Module 730: Sporadic False Positive Elimination Module 731 : Period setting module 733 : Window setting module by period 735: Density Calculation Module 737: False Detection Area Removal Module 750 : Algae Detection and Removal Module 751: Dominant Image Determination Module 753: Band-by-band Reflectance Value Comparison Module 755 : Recrystallization Module Th1: 1st threshold Th2: 2nd threshold Th3: Third threshold S1: Method for Removing Green Algae Pixels from Sargassum Candidates via Analysis of Geostationary Ocean Color Satellite Imagery

Claims

Claim 1 A preprocessing unit for preprocessing satellite imagery; a first determination unit for determining floating algae pixels by considering pixel-specific reflectance values ​​in the preprocessed satellite imagery; a second determination unit for determining the remaining pixels as Sargassum candidate pixels, excluding pixels determined as cloud regions among the floating algae pixels identified by the first determination unit through cloud region detection; and a false detection removal unit for automatically removing type-specific false detection pixels among the Sargassum candidate pixels; wherein the false detection removal unit includes an algae detection removal module for removing algae detection regions within the satellite imagery in which Sargassum candidate pixels are identified through the second determination unit; and the algae detection removal module includes a dominant image determination module for determining whether the image is Sargassum dominant or algae dominant by comparing the magnitudes of the reflectance values ​​of the green band wavelength and the red band wavelength for each of the Sargassum candidate pixels within the satellite imagery. A band-specific reflectance value comparison module that determines Sargassum pixels or green algae pixels by comparing the magnitudes of the reflectance values ​​of the green band wavelength and the red band wavelength of individual pixels identified as Sargassum candidates determined by Sargassum dominant images and green algae dominant images; A system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis, comprising: a recrystallization module that recrystallizes the suitability of a determined pixel when it is determined to be an algae pixel in a Sargassum dominant image, or conversely, when it is determined to be a Sargassum pixel in a Sargassum dominant image; wherein the recrystallization module determines the pixel as an algae pixel if the reflectance value of the green band wavelength minus the reflectance value of the red band wavelength is greater than or equal to a second threshold value when it is determined to be an algae pixel in a Sargassum dominant image, and determines it as a Sargassum pixel if it is less than the second threshold value; and when it is determined to be a Sargassum pixel in a Sargassum dominant image, determines the pixel as an algae pixel if the reflectance value of the red band wavelength minus the reflectance value of the green band wavelength is less than or equal to a third threshold value, and determines it as a Sargassum pixel if it is greater than the third threshold value. Claim 2 delete Claim 3 delete Claim 4 A system for removing green algae pixels within a Sargassum candidate group through geostationary satellite image analysis according to claim 1, wherein the band-specific reflectance value comparison module confirms a pixel as a Sargassum pixel when the reflectance value of the red band wavelength at a pixel identified as a Sargassum candidate group within a Sargassum dominant image is greater than the reflectance value of the green band wavelength. Claim 5 delete Claim 6 delete Claim 7 delete

Citation Information

Patent Citations

  • Method and system for detection of algal patches floating on sea surface using satellite imagery

    KR101737354B1

  • Extract and judge image of topography based on domain expansion

    KR1020080094503A

  • METHOD FOR DETECTING A SARGASSUM USING MULTISPECTRAL IMAGES LOADED ON MANNED and UNMANNED AERIAL VEHICLES AND SYSTEM THEREOF

    KR102579082B1