Remote sensing image quality evaluation method for shallow water area of island reef

Abstract

The invention relates to a remote sensing image quality evaluation method for an island shallow water area. The method comprises the following steps: separating an island underwater area from a land area by adopting a normalized water body index; extracting the shallow water area of the island reef by adopting relative water transparency and suspended solid total turbidity; constructing a cloud index, a cloud shadow index and a flare index to evaluate the cloud, shadow and flare conditions of the island shallow water region, and finally obtaining a plurality of special indexes; and 4, evaluating the image quality of the shallow water region of the island reef by using a special index and a general index (standard deviation, information entropy and definition). A remote sensing physical model and a traditional image quality evaluation method are organically combined, and the problem that an existing method is insufficient in applicability in an island shallow water area is effectively solved. By establishing a targeted quality evaluation system, accurate evaluation of an effective shallow water area in a satellite image is realized, and an automatic and efficient image quality evaluation solution is provided for island remote sensing monitoring.

Description

Technical Field

[0001] The present invention relates to a remote sensing image quality evaluation method for island and reef shallow water areas, and in particular to a method for constructing special indicators and general indicators for remote sensing image quality evaluation in island and reef shallow water areas, as well as a remote sensing image quality automatic evaluation method based on the special indicators and general indicators. Background Art

[0002] Remote sensing image quality assessment technology is a crucial prerequisite for marine satellite remote sensing and ecological monitoring. Currently, the industry primarily employs two approaches to remote sensing image quality assessment for shallow-water areas of islands and reefs: objective evaluation methods based on traditional numerical calculations and intelligent evaluation methods based on deep learning.

[0003] In traditional numerical computing, researchers typically use global statistics (such as pixel mean, standard deviation, and signal-to-noise ratio) combined with spatial characteristics (information entropy and gradient clarity) to construct comprehensive evaluation systems. However, these methods have significant limitations: First, global statistical features cannot effectively characterize the local quality characteristics of key shallow water areas of interest, such as coral reefs and seagrass beds, which are of interest to satellite remote sensing. Second, existing indicator systems do not consider the attenuation effect of water optical properties on reflectance spectra, resulting in significant differences in evaluation results under different water environments.

[0004] The deep learning-based quality assessment method builds a convolutional neural network model and uses manually annotated quality labels for supervised training. However, this method suffers from two fundamental flaws: First, it relies on subjective experience. The labeling process relies on expert experience to grade the quality of image blocks, and the Kappa coefficient between different annotators is only 0.65. Second, it is plagued by computing resource constraints. The evaluation of a single high-resolution image takes as long as 3.7 minutes (using GPU acceleration), making it difficult to meet the real-time monitoring needs of large island and reef areas in the South China Sea.

[0005] Of particular note, the unique marine environment of shallow reef areas presents unique challenges for quality assessment: 1) Dynamically changing water transparency directly affects satellite visibility depth, and existing methods lack a mapping relationship between transparency parameters and quality scores; 2) Instantaneous water depth changes (±2m) caused by tidal effects alter the bottom reflectivity curve, but traditional metrics cannot capture these transient characteristics; and 3) The relationship between clouds, shadows, and remote sensing in satellite imagery and traditional image evaluation metrics is not effectively modeled. This results in a lack of relevance between existing assessment systems and satellite remote sensing monitoring missions for shallow waters and shallow seas.

[0006] Therefore, there is an urgent need to develop a remote sensing image quality assessment method for shallow-water island and reef areas. This method should possess the following core features: 1) it should be able to quantitatively reflect the impact of water environment parameters such as transparency and suspended matter concentration on the quality of shallow-water island and reef areas; 2) it should be able to quantitatively measure the impact of clouds, shadows, and flares in satellite remote sensing imagery on the quality of shallow-water island and reef areas; and 3) it should enable lightweight and automated assessment to adapt to edge computing environments. This will provide reliable data quality assurance for marine island and reef ecological monitoring. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to overcome the problem that the existing methods focus on quantifying the quality of the entire remote sensing image and cannot fully measure the signal quality of shallow water areas that satellite remote sensing is concerned about, which leads to the failure of existing quality evaluation methods. A remote sensing image quality evaluation method for island and reef shallow water areas is proposed. Compared with previous methods, the present invention fully considers the characteristics of island and reef shallow water areas.

[0008] In order to solve the above technical problems, the present invention proposes a remote sensing image quality evaluation method for island and reef shallow water areas, comprising the following steps: Step 1: Separate the original remote sensing image into land and water, and filter out the water image after land by calculation. The ratio of the area of ​​the original remote sensing image to the total area of ​​the original remote sensing image is used to obtain the first special index. S 1; Step 2: Construct relative water transparency (RWT) and total suspended solids turbidity (TSST) calculation models to extract shallow water images. , by calculating shallow water images Area and water body images The ratio of the area to obtain the second special index S 2; Step 3: Remove shallow water images The cloud noise and shadow noise in the image are removed to obtain a noise-free shallow water image. , calculate noise-free shallow water images Area and shallow water image The ratio of the area to obtain the third special index S 31 ; Step 4: Build the flare index GI and apply it to shallow water images. Perform flare removal to obtain a noise-free shallow water image after flare correction. Calculate the ratio of the difference between the image before and after correction and the image before correction to obtain the fourth special index. S 32 ; Step 5: Use standard deviation, information entropy, and clarity evaluation algorithms to calculate the noise level in shallow water areas of islands and reefs, and obtain three general standard deviation indicators: S 4. Information entropy index S5. Clarity index S 6; Step 6: Sort and score the indicators of multiple images to be evaluated in the study area according to their quality. The image to be evaluated with the highest sum of scores of all indicators is the image with the best quality.

[0009] The present invention first uses the Normalized Difference Water Index (NDWI) to separate underwater and land areas; then, uses the Relative Water Transparency (RWT) and Total Suspended Solids Turbidity (TSST) to extract areas of interest for coral reef bottom classification; then, uses the Cloud Index (CI), Cloud Shadow Index (CSI), and Glint Index (GI) to evaluate the cloud, shadow, and glare conditions in shallow water bottom areas; finally, uses the standard deviation, information entropy, and clarity as universal indicators to measure the image quality of the bottom area.

[0010] The remote sensing image quality evaluation method for island and reef shallow water areas proposed in this invention has the following benefits: 1. A method for separating land and water based on the Normalized Difference Water Mass Index (NDWI) and a shallow water area extraction technique for islands and reefs, combining relative water transparency (RWT) and total suspended solids turbidity (TSST), was proposed. This technique, for the first time, enables a localized quantitative assessment of the quality of shallow water satellite remote sensing signals, addressing the lack of sensitivity of traditional global indicators to key areas. 2. Integrate the Cloud Index (CI), Cloud Shadow Index (CSI), and Glare Index (GI) to form a satellite image interference evaluation system, establish a quantitative correlation model between shallow water area quality and atmospheric interference factors, and overcome the limitation of traditional methods that ignore the dynamic impact of clouds; 3. Through a hybrid architecture combining specialized metrics (RWT / TSST / CI / CSI / GI) with general metrics (standard deviation / information entropy / clarity), the system reduces single-image processing time to 45 seconds (CPU environment) while ensuring assessment accuracy. This represents a computational efficiency improvement of over five times compared to existing deep learning methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The present invention will be further described below with reference to the accompanying drawings.

[0012] Figure 1 This is an overall flow chart of an example of the present invention.

[0013] Figure 2This is the RWT shallow water extraction result of the example of the present invention: (a) original remote sensing image; (b) shallow water body (yellow area) extraction result.

[0014] Figure 3 This is the TSST shallow water extraction result of the present invention: (a) original remote sensing image; (b) shallow water body (orange area) extraction result.

[0015] Figure 4 This is the CI cloud extraction result of the example of the present invention.

[0016] Figure 5 This is the CSI shadow extraction result of the example of the present invention.

[0017] Figure 6 These are the experimental data for Sentinel-2 image quality evaluation. The imaging times are (a) 2020.01.19, (b) 2020.03.31, (c) 2020.04.20, and (d) 2021.07.27. DETAILED DESCRIPTION

[0018] The present invention is described in detail below with reference to the accompanying drawings to make the technical route and operation steps of the present invention clearer. Figure 1 .

[0019] Step 1: First dedicated indicator S 1 Calculation: Assume the original remote sensing image is I , using the normalized water index NDWI Separate land and water, normalized water index of pixels NDWI Pixels that meet any of the following two formulas are land pixels and will be removed: Where, Indicates the radiation brightness value in the near-infrared band, Indicates the radiation brightness value of shortwave infrared 1, Indicates the radiance value of the green light band.

[0020] In this embodiment, the image after water and land separation can be expressed as , where the land area is assigned a value of 0. So far, after filtering out the land part, the first dedicated index for evaluating the image quality score can be obtained. S 1. The first dedicated indicator S 1 can be passed The ratio of the area of ​​the original remote sensing image to the total area of ​​the original remote sensing image can also be calculated using the following formula: The total number of pixels in the image whose pixel value is not 0. NDWI After removing the land, the next step is to use relative water transparency (RWT) and total suspended solids turbidity (TSST) to extract shallow water areas of interest for bottom sediment classification in coral reefs.

[0021] Step 2: Construct the relative water transparency (RWT) and total suspended solids turbidity (TSST) calculation model, and use the automatic threshold segmentation method to extract the shallow water areas of interest for island and reef satellite remote sensing monitoring. , by calculating the shallow water area Area and water body images The ratio of the area to obtain the second special index S 2.

[0022] In this step, according to the spectral transmission characteristics, the relative water transparency RWT of the pixel satisfies the following formula, which is the deep water pixel and is removed (the deep water pixel is assigned a value of 0) to obtain the relative water transparency shallow water image : Where, e is the natural logarithm, Indicates the radiation brightness value of the red light band, represents the radiation brightness value of the green light band, r represents the composite brightness index of the red light band and the green light band, T wt is the automatic separation threshold, T wt Calculated by the following formula: Where, Indicates the minimum value of relative water transparency RWT, Indicates the average value of relative water transparency RWT. is the first adjustment coefficient, generally selected from {1 / 3, 1 / 2, 2 / 3, 3 / 4, 4 / 5}.

[0023] The results of deep and shallow water separation after segmentation by relative water transparency threshold are shown in Figure 2 , Figure 2 (a) is the water image before separation. Figure 2 (b) is the result after segmentation by relative water transparency threshold. In the figure, the yellow area is shallow water.

[0024] Total suspended solids turbidity (TSST) is used to assess the clarity and quality of water bodies. To a certain extent, it can extract shallow water areas of interest for bottom sediment classification. Pixels whose total suspended solids turbidity (TSST) satisfies any of the following two formulas are deep water pixels and are removed (deep water pixels are assigned a value of 0) to obtain a shallow water image of total suspended solids turbidity. : Where, R Indicates the radiation brightness value of the red light band, B Indicates the radiant brightness value of the blue light band, T tsst is the automatic segmentation threshold of total suspended solid turbidity TSST, which is calculated by the following formula: Where, Indicates the minimum value of total suspended solids turbidity TSST, It represents the average value of total suspended solid turbidity TSST. is the second adjustment coefficient, generally selected from {1 / 4, 1 / 3, 1 / 2, 2 / 3, 3 / 4}.

[0025] The results of deep and shallow water separation after segmentation by total suspended solids turbidity threshold are shown in Figure 3 , Figure 3 (a) is the water image before separation. Figure 3 (b) is the result after segmentation by relative water transparency threshold. In the figure, the orange area is shallow water.

[0026] Relative water transparency shallow water image Shallow water imaging with total suspended solids turbidity After merging, the shallow water image after deep and shallow water separation in this step can be obtained, which is expressed as : At this point, after extracting the shallow water area of ​​interest for bottom classification, the second dedicated indicator for evaluating image quality is obtained. S 2. Second dedicated indicator S 2 can pass shallow water image Area and water body images The ratio of the areas can also be calculated using the following formula: The total number of pixels in the image whose pixel value is not 0.

[0027] It is worth mentioning that RWT and TSST reflect the transparency and biological parameters of the water body, so they also have a certain filtering effect on clouds. However, they cannot evaluate thin clouds, shallow shadows, etc. in the area of ​​interest. Therefore, after using RWT and TSST to extract the shallow water area of ​​concern for bottom sediment classification, CI, CSI, and GI are continued to be used to evaluate the environmental noise conditions in the area.

[0028] Step 3: Filter out shallow water images Remove the cloud and shadow noise in the image to obtain a noise-free shallow water image. , calculate noise-free shallow water images Area and shallow water image The ratio of the area to obtain the third special index S 31 .

[0029] Among them, noise-free shallow water images Obtained by the following calculation: Where, Indicates shallow water image Cloud images in Indicates shallow water image Shadow image in .

[0030] The following describes in detail the extraction methods of cloud index (cloud noise) and shadow index (shadow noise).

[0031] Most low- to medium-resolution multispectral / hyperspectral remote sensing sensors (such as Sentinel-2 and Landsat series) include both near-infrared (NIR) and short-wavelength infrared (SWIR) spectral channels, but high-resolution sensors (such as Gaofen-2) typically only have NIR spectral channels. Depending on whether the SWIR band is included, the CI (Cloud Index) is calculated in two ways: The CI1 index is used to measure the similarity of reflectance characteristics in the visible and infrared bands. Generally speaking, since clouds exhibit similar reflectance characteristics in the visible and infrared bands, the CI1 index usually fluctuates within a narrow range very close to 1. Indicates the radiation brightness value in the near-infrared band, Indicates the short-wave infrared radiation brightness value, Indicates the radiant brightness value of the green light band, Indicates the radiation brightness value of the red light band, Indicates the radiance value of the blue light band.

[0032] The CI2 index, which is the average value of the relevant band spectrum, is used to describe the brightness characteristics of clouds. Indicates the radiation brightness value in the near-infrared band, Indicates the radiation brightness value of shortwave infrared 1, Indicates the radiation brightness value of shortwave infrared 2, Indicates the radiant brightness value of the green light band, Indicates the radiation brightness value of the red light band, Indicates the radiance value of the blue light band.

[0033] For remote sensing images from different satellites, the index cloud layer can be effectively separated by any of the following formulas: Y1 is a smaller threshold, and Y2 is a larger threshold. The first parameter, Y1, can generally be selected from the set {0.01, 0.1, 1, 10, 100} and can be fine-tuned for specific scenarios through experimentation. Parameter Y2 is set to a larger value. Taking into account the reflective properties of different coral reef substrates, in practical applications, Y2 can be adaptively determined using the following method: in, Indicates index The average value of Indicates index The maximum value of is the adjustment coefficient, usually selected from {1 / 10, 1 / 9, 1 / 8, 1 / 7, 1 / 6, 1 / 5, 1 / 4, 1 / 3, 1 / 2}. Figure 4 The results of cloud extraction using the CI index from Sentinel-2 remote sensing images of a certain island reef (red area) are shown.

[0034] The calculation of the CSI index (shadow index) is similar to that of the CI index (cloud index). Depending on whether the SWIR band is included, the SI index can be constructed using the following two formulas: Among them, the CSI index is designed for NIR band and The mean of the band, or just the NIR band, is used to describe the reflective properties of cloud shadows (shadows in island and reef images are generally cloud shadows) at longer wavelengths. However, water generally exhibits similar reflective properties within these wavelength ranges. Therefore, in order to accurately identify cloud shadows projected on the water surface and eliminate water interference, additional constraints need to be introduced. Given that water has a higher reflectivity at shorter wavelengths, especially in the blue light band, this characteristic of the blue light band can be used to eliminate water interference. Specifically, the following two formulas must be satisfied at the same time: In the formula, the parameter Y3 is generally set to a smaller value. The calculation method is as follows: in, express The minimum value of the index, express The average value of the index, is the adjustment coefficient, which is generally selected from {1 / 4, 1 / 3, 1 / 2, 2 / 3, 3 / 4}. The fourth parameter Y4 can be determined in practical applications as follows: in, Indicates the minimum value of the blue light band radiation brightness value, Represents the average value of the blue light band radiation brightness value, is the adjustment coefficient, generally selected from {1 / 2, 2 / 3, 3 / 4, 4 / 5, 5 / 6}.

[0035] Figure 5 The image shows the results of cloud shadow extraction (purple area) from the Sentinel-2 remote sensing image of a certain island reef using CSI.

[0036] The noise-free shallow water image after removing clouds and shadows can be expressed as : At this point, the third special index can be calculated according to the following formula S 31 : The third special indicator S 31 Noise-free shallow water images Area and shallow water image The ratio of the areas can also be calculated using the following formula: Where, The total number of pixels in the image whose pixel value is not 0.

[0037] Step 4: Build the flare index GI and apply it to shallow water images. Perform flare removal to obtain a noise-free shallow water image after flare correction. Calculate the ratio of the difference between the image before and after correction and the image before correction to obtain the fourth special index. S 32 .

[0038] In this embodiment, the flare index GI calculation adopts an improved flare removal algorithm. Specifically, in shallow water images Select the points in the sample area, take the brightness of the near-infrared band as the X-axis and the brightness of the visible light band as the Y-axis, perform linear regression on all pixels in each visible light band, and calculate their linear relationship. i The slope of the linear regression b i , the flare removal formula is: Where, Indicates the band after flare removal i The radiance value, Indicates band i The radiance value, express The radiance value of the pixel after band operation; express The minimum radiance value after band calculation (which can be considered as the radiance value of the pixel without the influence of solar flare) Indicates band i The slope of the linear regression. The image after flare correction is obtained. By evaluating the difference between the image before and after correction, a special index can be obtained. S 32 : Where, Indicates shallow water image Band i The average value of the ratio of the difference in radiance before and after flare correction to the radiance value before correction for all pixels.

[0039] S 31 With S 32 The special index S3 can be obtained by adding them together. Using S1, S2, and S3, the special index score S of the remote sensing image to be screened can be obtained. SP = S1+ S2+ S3, S SP The higher the value, the better the image quality.

[0040] Step 5: Use standard deviation, information entropy, and clarity evaluation algorithms to calculate the noise level in shallow water areas of islands and reefs, and obtain three general standard deviation indicators: S 4. Information entropy index S 5. Clarity index S 6.

[0041] The calculation of the standard deviation index S4 can be used to measure the degree of change in the pixel values ​​in the image. If the standard deviation is large, it means that the pixel values ​​in the image are dispersed and have large changes. If the standard deviation is small, it means that the pixel values ​​in the image are concentrated and have small changes. For remote sensing images, the larger the standard deviation, the worse the image quality (poor), and the smaller the standard deviation, the better the image quality (excellent). Suppose image I S The total number of pixels is N The average radiation brightness is , For pixels i The radiation brightness value, then the standard deviation : Information entropy index S 5 is used to measure the uncertainty or randomness of the pixel values ​​in an image. If the pixel values ​​in an image are evenly distributed, the information entropy is high; if the pixel values ​​in an image are concentrated, the information entropy is low. For remote sensing images, the higher the information entropy, the better the image quality (excellent), and the lower the information entropy, the worse the image quality (poor). For image I S , L represents the number of gray levels of the image, then the information entropy It can be expressed as: Where, Represents pixels i The radiance value The probability of appearing in the image, Represents pixels i The radiance value.

[0042] The image clarity index S6 calculation methods include Brenner, Tenengrad, and SMD2. Among them, the clarity obtained by Brenner has attracted widespread attention because it is consistent with human subjective perception. Therefore, Brenner is selected to calculate the clarity of remote sensing images. The Brenner gradient function is the simplest gradient evaluation function. It simply calculates the square of the grayscale difference between two adjacent pixels. The function is defined as follows: in, Representing an image Corresponding pixel Gray value.

[0043] For remote sensing images, the higher the clarity, the better (excellent) the image quality, and the lower the clarity, the worse (poor) the image quality.

[0044] Step 6: Sort and score the indicators of multiple images to be evaluated in the study area according to their quality. The image to be evaluated with the highest sum of scores of all indicators is the image with the best quality.

[0045] In this step, based on the constructed special indicators and general indicators, the overall screening process of high-quality remote sensing images is as follows: Calculate the image to be evaluated. S 1. S 2. And S 3, and sort and assign values ​​according to the scores, assign three scores [N, 1] to each candidate image from high to low; calculate the image to be evaluated S 4. S 5. and S 6, and sort and assign values ​​according to the scores, and assign three scores [N, 1] to each candidate image according to the scores. If the smaller the standard deviation, the less image noise, the standard deviation index S The smallest image is assigned a score of N, and the largest image is assigned a score of 1. The six scores of the images to be evaluated are summed to obtain the final score. The images to be evaluated are sorted by their final scores, with the image with the highest score being the best.

[0046] Verification Example In order to verify the accuracy and reliability of the method of the present invention, the following example is used for further explanation.

[0047] Taking a certain island reef as an example, the proposed image screening method was verified using four Sentinel-2 images with different imaging times and obvious quality differences. Figure 6 , Figure 6 The imaging time of (a) is 2020.01.19, Figure 6 The imaging time of (b) is 2020.03.3, Figure 6 The imaging time of (c) is 2020.04.20. Figure 6 (d) Image imaging time: 2021.07.27.

[0048] Table 1 Image quality evaluation results It can be seen from Table 1 that Figure 6 (b) has the best image quality. Figure 6 (c) Image quality is second to none, Figure 6 (d) The image quality is the worst. Compared with the other three images, Figure 6 (b) The image has the highest effective bottom area, and there are no obvious clouds, shadows, or flares. The overall situation is good, so it received the highest score of 24 points. Figure 6 (a)'s dedicated index ranked second (8 points), higher than the image Figure 6 (c) 7 points, but from the general indicators, Figure 6 (c) The image scored 9 points. Figure 6 (a) The image is only worth 5 points, so Figure 6 (c) Image is ranked second. Figure 6 (d) Image ranked last in almost all indicators, with a comprehensive score of only 7 points, which is Figure 6 (a) Compared with the image, Figure 6 (d) The cloud cover in the image is less, but the clouds in this image mainly cover the shallow water area that is the focus of bottom classification. Figure 6 (a) Although the image has the most clouds, they are less distributed in the shallow water area. The above experiments fully demonstrate the effectiveness of the image quality evaluation method proposed in this paper.

[0049] The remote sensing image quality evaluation method for island and reef shallow water areas of the present invention is not limited to the specific technical solutions described in the above embodiments. All technical solutions formed by equivalent replacement are within the protection scope required by the present invention.

Claims

1. A remote sensing image quality assessment method for shallow water areas of islands and reefs, comprising the following steps: Step 1: Separate the original remote sensing image into land and water, and filter out the water image after land by calculation. The ratio of the area of ​​the original remote sensing image to the total area of ​​the original remote sensing image is used to obtain the first special index. S 1; Step 2: Construct relative water transparency (RWT) and total suspended solids turbidity (TSST) calculation models to extract shallow water images. , by calculating shallow water images Area and water body images The ratio of the area to obtain the second special index S 2; Step 3: Remove shallow water images The cloud noise and shadow noise in the image are removed to obtain a noise-free shallow water image. , calculate noise-free shallow water images Area and shallow water image The ratio of the area to obtain the third special index S 31 ; Step 4: Build the flare index GI and apply it to shallow water images. Perform flare removal to obtain a noise-free shallow water image after flare correction. Calculate the ratio of the difference between the image before and after correction and the image before correction to obtain the fourth special index. S 32 ; Step 5: Use standard deviation, information entropy, and clarity evaluation algorithms to calculate the noise level in shallow water areas of islands and reefs, and obtain three general standard deviation indicators: S 4. Information entropy index S 5. Clarity index S 6; Step 6: Sort and score the indicators of multiple images to be evaluated in the study area according to their quality. The image to be evaluated with the highest sum of scores of all indicators is the image with the best quality.

2. The remote sensing image quality assessment method for shallow water areas of islands and reefs according to claim 1 is characterized by: In step 1, the normalized water index NDWI Separate land and water from the original remote sensing image, and normalize the water index of the pixel NDWI Pixels that meet any of the following two formulas are land pixels and will be removed: Where, Indicates the radiation brightness value in the near-infrared band, Indicates the radiation brightness value of shortwave infrared 1, Indicates the radiance value of the green light band.

3. The remote sensing image quality assessment method for shallow water areas of islands and reefs according to claim 1 is characterized by: In step 2, the pixel whose relative water transparency RWT satisfies the following formula is a deep water pixel and is removed to obtain a shallow water image with relative water transparency: : Where, e is the natural logarithm, Indicates the radiation brightness value of the red light band, represents the radiation brightness value of the green light band, r represents the composite brightness index of the red light band and the green light band, T wt is the automatic separation threshold of relative water transparency RWT, which is calculated by the following formula: Where, Indicates the minimum value of relative water transparency RWT, Represents the average value of relative water transparency RWT, is the first adjustment coefficient, .

4. The remote sensing image quality assessment method for shallow water areas of islands and reefs according to claim 1 is characterized by: In step 2, the total suspended solid turbidity TSST of the pixel that satisfies any of the following two formulas is a deep water pixel and is removed to obtain the shallow water image of the total suspended solid turbidity : Where, Indicates the radiation brightness value of the red light band, Indicates the radiant brightness value of the blue light band, T tsst is the automatic segmentation threshold of total suspended solid turbidity TSST, which is calculated by the following formula: Where, Indicates the minimum value of total suspended solids turbidity TSST, It represents the average value of total suspended solid turbidity TSST. is the second adjustment coefficient, .

5. The remote sensing image quality assessment method for shallow water areas of islands and reefs according to claim 1 is characterized by: In step 3, the cloud index CI and shadow index CSI are constructed, and the automatic threshold method is used to remove shallow water images. Cloud noise and shadow noise in the image, noise-free shallow water image Obtained by the following calculation: Where, Indicates shallow water image Cloud images in Indicates shallow water image Shadow image in .

6. The remote sensing image quality assessment method for shallow water areas of islands and reefs according to claim 5, characterized in that: The third dedicated indicator S 31 The calculation is as follows: Where, The total number of pixels in the image whose pixel value is not 0.

7. The remote sensing image quality assessment method for shallow water areas of islands and reefs according to claim 1 is characterized by: In step 4, GI calculation is first performed on the shallow water image. Select the points in the sample area, take the brightness of the near-infrared band as the X-axis and the brightness of the visible light band as the Y-axis, perform linear regression on all pixels in each visible light band, calculate the linear relationship, and obtain the band i The slope of the linear regression, flare removal formula is: Where, Indicates the band after flare removal i The radiance value, Indicates band i The radiance value, express The radiance value of the pixel after band operation; express The minimum radiance value after band operation, Indicates band i The slope of the linear regression.

8. The remote sensing image quality assessment method for shallow water areas of islands and reefs according to claim 7 is characterized by: Fourth Special Indicator S 32 Calculated by the following formula: Where, Indicates shallow water image Band i The average value of the ratio of the difference in radiance before and after flare correction to the radiance value before correction for all pixels.

9. The remote sensing image quality assessment method for shallow water areas of islands and reefs according to claim 1, characterized in that: The third dedicated indicator S 31 With the fourth dedicated indicator S 32 The sum is the fifth special index S 3. In step 6, each image to be evaluated is evaluated according to the first special index S 1. Second dedicated indicator S 2. The fifth special indicator S 3. Standard Deviation Indicator S 4. Information entropy index S 5 and clarity index S 6, and assign three scores [1, N] to each candidate image from best to worst; add up the scores of the six indicators of the image to be evaluated to get the final score, and sort the candidate images according to the final score. The image with the highest score is the best image.

Images