Sea ice remote sensing identification method based on ultraviolet reflection difference

Abstract

The invention relates to a sea ice remote sensing identification method based on ultraviolet reflection difference, and the method achieves the quick and accurate identification of sea ice in a multi-source remote sensing image through the unique optical signal difference of sea ice, cloud and other confusion targets under the reflection of an ultraviolet band. And estimating the sea ice concentration and the sea ice coverage area of the sea ice image after the cloud mask. According to the method, timely and efficient guidance information can be provided, the used multi-source remote sensing image only needs to be subjected to radiometric calibration and atmospheric correction to obtain an optical band apparent reflectivity and a thermal infrared band brightness temperature image, and sea ice can be accurately identified and extracted; compared with a traditional optical observation means, the method is suitable for the situation that an optical remote sensing image is prone to cloud coverage interference, and cloud interference information can be recognized and removed at the same time; the method can meet the requirements of sea ice business monitoring and improve the working efficiency of sea ice observation.

Description

Technical Field

[0001] The invention relates to the technical field of satellite remote sensing polar sea ice monitoring, and in particular to a sea ice remote sensing identification method based on ultraviolet reflection difference. Background Art

[0002] Sea ice is a key target for marine environmental monitoring. Its range and thickness vary greatly, and it can cover about 10% of the Earth's oceans. Sea ice plays an important role in the global system and has been integrated into climate models. It can have a profound impact on polar and global climate, oceans, and ecosystems. In addition, sea ice will also affect human scientific research and production activities in the polar regions. It is an important factor threatening human shipping and navigation in polar waters, as well as the safety of offshore oil platform exploitation. Research on sea ice changes is mainly carried out through its area, coverage, density, thickness, surface temperature, etc.

[0003] The development of satellite remote sensing technology has opened up new ways for large-scale monitoring of sea ice. It can cover a vast sea area in a short period of time and obtain key information such as the distribution range and thickness of sea ice in a timely manner. Among them, satellite optical remote sensing provides a good technical support for the dynamic monitoring of sea ice. For example, China's Haiyang-1C / D satellite (HY-1C / D) and the United States' Moderate-resolution Imaging Spectroradiometer (MODIS) have been widely used in the observation and research of polar sea ice. Optical remote sensing data can provide multi-band and even hyperspectral information. Based on the reflectivity difference between sea ice and background seawater, the band threshold method can be used to effectively identify sea ice; and for sea ice monitoring in some small areas or specific areas, the use of high spatial resolution optical remote sensing data can achieve more detailed observations. In the process of sea ice observation, sea ice concentration (SIC) and sea ice extent are important parameters to describe sea ice, which are defined as the proportion of sea ice coverage per unit area and the total amount of sea ice coverage in a certain area, respectively. The sea ice extent can be calculated from the sea ice concentration.

[0004] Optical remote sensing data is used to carry out remote sensing monitoring of sea ice, providing sea ice density and sea ice range. However, due to differences in the band settings and signal-to-noise ratio of different optical satellite payloads, current optical remote sensing monitoring of sea ice is mainly interfered by clouds, both of which appear as high reflectivity features in optical remote sensing images. The new generation of multi-band remote sensing image data is equipped with ultraviolet bands. Sea ice, clouds, sea water, etc. have unique scattering characteristics in their respective reflectivity spectra. These features appear in the form of peaks and valleys, and their positions and values ​​can be used to remove interference information and extract sea ice information. Therefore, the accurate identification and extraction of sea ice and its key information is of great practical significance for sea ice monitoring. Summary of the invention

[0005] The technical problem to be solved by the present invention is: in view of the fact that optical remote sensing identification of sea ice is easily interfered by confusing targets such as clouds, classification and identification are carried out based on the reflection difference of sea ice and clouds on incident light in the ultraviolet band, and the sea ice density and sea ice coverage area are estimated, which is of great significance for sea ice monitoring.

[0006] In order to solve the above technical problems, the present invention provides a sea ice remote sensing identification method based on ultraviolet reflectance difference, comprising the following steps:

[0007] Step 1: Data preprocessing

[0008] Preprocess multi-source remote sensing image data and convert optical band data into apparent reflectance ρ TOA Data, the optical band includes ultraviolet band, visible light band and near infrared band, and the thermal infrared band data is converted into brightness temperature data BT;

[0009] Step 2: Identify the area of ​​interest

[0010] For the image data preprocessed in step 1, use the vector file to crop out the area of ​​interest containing sea ice;

[0011] Step 3: Spectral information enhancement

[0012] For the apparent reflectance data ρ TOA Perform background water spectrum difference calculation, subtract background seawater spectrum pixel by pixel, and weaken the influence of background water color; normalize the background seawater difference data to obtain normalized difference reflectance data ρ' TOA ;

[0013] Step 4: Feature band detection

[0014] Select multiple typical spectral data of typical targets as training samples. The typical spectral data refers to the characteristic spectral data obtained by the spectrum determined as the designated target through the above steps 1-3. Take the average value of the characteristic spectral data of the training sample of each target, and calculate the contrast C between the spectral characteristic data according to the following formula:

[0015]

[0016] In the formula, ρ TOA,UV / RED is the apparent reflectance value of the typical target characteristic spectrum in the ultraviolet band or red light band, ρ TOA,VNIR is the apparent reflectance value of the characteristic spectrum of a typical target in the visible-near infrared band; the typical targets include sea ice, cirrus, cumulus and sea water; the two band combinations with the largest contrast C are selected for sea ice extraction and confusing target removal based on ultraviolet reflectance difference, and the final selected band combinations are: ultraviolet band and red light band combination, and ultraviolet band and near infrared band combination; step 5, sea ice recognition based on ultraviolet band reflectance difference is performed on the image, and the specific steps are as follows:

[0017] 1) For the apparent reflectance data ρ TOA The near infrared band grayscale image ρ TOA,NIR , using the threshold ε 1 Perform threshold segmentation to obtain background seawater pixels and other pixels to be segmented. The specific formula is as follows:

[0018]

[0019] In the water_mask image, pixels with a value of 1 represent seawater, and pixels with a value of 0 represent non-seawater;

[0020] 2) Use the Water_mask image as a mask to normalize the difference reflectance data ρ' TOA Operation is used to remove seawater pixels. For the normalized difference reflectance data ρ' of the removed seawater pixels TOA , use the ultraviolet band, green band, and red band to construct the corresponding index, and obtain the brightness temperature image data BT at the same time, using the threshold ε 2 , threshold ε 3 and threshold ε 4 Perform threshold segmentation to obtain cirrus cloud pixels. The specific formula is as follows:

[0021]

[0022] Where, T 1 =ρ' TOA,RED -ρ' TOA,GREEN , ρ' TOA,UVIt is the normalized difference reflectance data of the ultraviolet band. The pixels with a value of 1 in the Cirrus_mask image represent cirrus clouds, and the pixels with a value of 0 represent non-cirrus clouds;

[0023] 3) Use the Cirrus_mask image as a mask to normalize the difference reflectivity data ρ' TOA The operation is used to remove the seawater and cirrus pixels. For the normalized difference reflectance data ρ' after removing the seawater and cirrus pixels TOA , using the red light band and the near infrared band to construct the corresponding index, and at the same time obtain the brightness temperature image data BT, using the threshold ε 4 and threshold ε 5 Threshold segmentation is performed to obtain cumulus pixels and sea ice pixels. The specific formula is as follows:

[0024]

[0025] Among them, T 2 =ρ' TOA,NIR / ρ' TOA,RED , ρ' TOA,NIR is the normalized difference reflectance data in the near-infrared band, ρ' TOA,RED It is the normalized difference reflectance data of the red light band. The pixels with a value of 1 in the Cumulus_mask image represent cumulus clouds, and the pixels with a value of 0 represent non-cumulus clouds, which is the result of sea ice pixel extraction.

[0026] Based on the sea ice pixel extraction results in step 5, the most frequent apparent reflectance ρ is obtained by statistics. TOA The value is used as the apparent reflectance value of the pure sea ice pixel, and combined with the apparent reflectance value of the sea water pixel, the sea ice concentration SIC (Sea ice concentration) is calculated; according to the sea ice pixel area and the sea ice concentration, the sea ice coverage area S is calculated. ice (Sea ice extent).

[0027] By utilizing the unique optical signal differences of confusing targets such as sea ice and clouds under ultraviolet reflection, rapid and accurate identification of sea ice in multi-source remote sensing images is achieved, and the sea ice density and sea ice coverage area are estimated for the sea ice images after cloud masking. The present invention can provide timely and efficient guidance information, and the multi-source remote sensing images used only need to be calibrated by radiation and corrected by the atmosphere to obtain the apparent reflectivity of the optical band and the brightness temperature image of the thermal infrared band, so that the sea ice can be accurately identified and extracted; compared with traditional optical observation methods, the present invention is suitable for situations where optical remote sensing images are easily interfered by cloud coverage, and cloud interference information can be identified and eliminated at the same time (clouds are similar to sea ice in visual features and spectral morphological features, and examples are given in the specification of the present invention to illustrate); this method can meet the needs of sea ice business monitoring and improve the work efficiency of sea ice observation. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram of the process of the present invention.

[0029] Figure 2 This is a HY-1C / D true color composite remote sensing image of sea ice, clouds and background seawater in the area of ​​interest.

[0030] Figure 3-a This is the apparent reflectance image of the area of ​​interest in the 355nm ultraviolet band.

[0031] Figure 3-b It is the brightness temperature image of the area of ​​interest in the 1080nm thermal infrared band.

[0032] Figure 4 Apparent reflectivity of sea ice, clouds and background seawater, background seawater difference, and normalized difference reflectivity spectra.

[0033] Figure 5 Schematic diagram of the contrast index for training-recognition of the present invention.

[0034] Figure 6 A schematic diagram of setting the target detection classification threshold of the present invention.

[0035] Figure 7 This is a diagram of sea ice and cloud identification results according to the present invention.

[0036] Figure 8 This is a diagram showing the estimation results of sea ice density and sea ice coverage area according to the present invention. DETAILED DESCRIPTION

[0037] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more definite definition of the protection scope of the present invention.

[0038] This embodiment is applied to several scenes of HY-1C / D satellite multi-band remote sensing image data from June to August 2022, covering the Chukchi Sea and Beaufort Sea in the Arctic. The images include 10 optical bands of 355nm, 385nm, 412nm, 443nm, 490nm, 520nm, 565nm, 670nm, 750nm, and 865nm, and 2 thermal infrared bands of 1080nm and 1200nm, with a pixel spatial resolution of 1100m.

[0039] like Figure 1 FIG. 1 is a flow chart of a method for remote sensing identification and density estimation of sea ice based on ultraviolet reflectance difference according to an embodiment of the present invention, which includes the following steps:

[0040] Step 1: Data preprocessing

[0041] Preprocess the HY-1C / D satellite multi-band remote sensing image data, including geometric correction, radiation calibration, atmospheric correction and smoothing filtering, and convert the optical band data into apparent reflectance ρ TOA Data, convert thermal infrared band data into brightness temperature BT data.

[0042] Step 2: Identify the area of ​​interest

[0043] The regions of interest selected in this embodiment are the Chukchi Sea and the Beaufort Sea in the Arctic. For the sea ice distribution area, the region of interest containing the sea ice is delineated on the image pre-processed in step 1 using the vector file of the study area.

[0044] In this embodiment, Figure 2 , 3 are the HY-1C / D satellite true color synthetic images of the 6 interest areas after preprocessing in steps 1 and 2, and the apparent reflectance ρ of the 355nm ultraviolet band TOA Grayscale image and 1080nm thermal infrared band brightness temperature BT grayscale image. Sea ice and clouds appear as similar visual features in the image, which poses certain challenges to the identification and extraction of sea ice.

[0045] Step 3: Spectral information enhancement

[0046] The apparent reflectivity ρ of the research target (sea ice), typical interference targets (cirrus, cumulus) and background seawater is detected. TOA The spectral curve is averaged, such as Figure 4 -a to Figure 4 -d, build a spectrum set.

[0047] P TOA The background water spectrum difference is calculated by subtracting the background seawater spectrum pixel by pixel to weaken the background water color shadow. Figure 4 -e shows the background seawater difference spectrum; the difference spectrum is normalized, and the result is as follows Figure 4 The above steps can effectively eliminate the difference in reflectivity values ​​caused by different regions and radiation energy, and improve the uniformity of the spectral morphology of the same target and the difference of the spectral morphology of different targets.

[0048] The apparent reflectivity ρ of the target is calculated using the following formula: TOA Subtract the background seawater value ρ TOA,seawater After normalization, the normalized difference reflectivity data ρ' is obtained. TOA,i , to achieve the effect of target enhancement:

[0049] a i =ρ TOA,i -ρ TOA,seawater i=1,2,3,…,8

[0050] ρ' TOA,i =(a i -a i,min ) / (a i,max -a i,min )

[0051] Where i is the band number of the optical band, a i is the background seawater difference of band i, a i,max and a i,min are the maximum and minimum values ​​in the difference spectrum curve of band i, respectively. Finally, the normalized difference reflectance data ρ' of band i is obtained. TOA,i .

[0052] like Figure 4 As shown, the original ρ of sea ice and cirrus and cumulus clouds TOA The spectral morphology is similar, and the above processing steps can increase the spectral feature differences of different targets and effectively distinguish them.

[0053] Step 4: Feature band detection

[0054] In this embodiment, all targets except sea ice are regarded as interference and background targets. The purpose of characteristic band detection is to suppress other target signals while enhancing sea ice signals. Multiple typical spectral data of typical targets (sea ice, cirrus, cumulus, sea water) in step 3 are selected as training samples, and the characteristic spectral data of the training samples of each target are averaged. The contrast index C is used to quantitatively evaluate the band combination effect. The calculation formula of contrast C is as follows:

[0055]

[0056] In the formula, ρ TOA,355 / 670nm is the apparent reflectance value of the typical target characteristic spectrum in the 355nm ultraviolet or 670nm red light band (the typical target characteristic spectrum has the maximum apparent reflectance in the above bands), ρ TOA,VNIR It is the apparent reflectance value of the typical target characteristic spectrum in the visible light-near infrared band (412nm-865nm).

[0057] The calculation results are as follows Figure 5 -e, the contrast C of the combination of 355nm ultraviolet band, 670nm red light band and 865nm near infrared band is the largest. The above band combination is used to construct a sea ice and cloud classification algorithm based on ultraviolet reflectance difference.

[0058] Step 5: Identify sea ice and classify targets based on images

[0059] Sea ice pixel identification and extraction is based on the difference in ultraviolet band reflectance. The specific steps are as follows: Figure 6 As shown:

[0060] 1) The background seawater has strong absorption characteristics in the near-infrared band, so for the near-infrared band grayscale image ρ TOA,865nm , set the threshold ε 1 =0.05 vs. TOA,865nm The image is segmented by threshold and the seawater pixel extraction result is obtained. The formula is as follows:

[0061]

[0062] In the water_mask image, pixels with a value of 1 represent seawater, and pixels with a value of 0 represent non-seawater.

[0063] 2) Use the Water_mask image as a mask to normalize the difference reflectance data ρ' TOA Operation is used to remove seawater pixels. For the normalized difference reflectance data ρ' of the removed seawater pixels TOA , the image contains sea ice, cirrus, and cumulus pixels. Figure 4 From the spectrum, we can see that cirrus clouds show strong scattering characteristics in the 355nm ultraviolet band and have a small reflection peak in the 670nm red band (which can be determined by the difference between the red and green bands T 1 =ρ' TOA,670nm -ρ' TOA,565nm In addition, compared with sea ice, the brightness temperature BT of high-altitude cirrus clouds in the 1080nm thermal infrared band is significantly lower. Therefore, using the above characteristic bands, the threshold ε is set 2 =0.25,ε 3 =0.5,ε 4 =266.5K for threshold segmentation, and the cirrus pixel extraction result is obtained. The formula is as follows:

[0064]

[0065] In the Cirrus_mask image, pixels with a value of 1 represent cirrus clouds, and pixels with a value of 0 represent non-cirrus clouds.

[0066] 3) Use the Cirrus_mask image as a mask to normalize the difference reflectivity data ρ' TOA The operation is used to remove the seawater and cirrus pixels. For the normalized difference reflectance data ρ' after removing the seawater and cirrus pixels TOA , the image contains sea ice and cumulus pixels. Figure 4 From the mid-spectrum, we can see that the spectral variation trend of cumulus clouds from red light to near infrared is relatively gentle (as can be seen from the ratio of near infrared to red light band T 2 =ρ' TOA,865nm / ρ' TOA,670nmIn addition, compared with sea ice, the brightness temperature BT of high-altitude cumulus clouds in the 1080nm thermal infrared band is significantly lower. Therefore, using the above characteristic bands, the threshold ε is set 4 =266.5K, ε 5 =0.75 for threshold segmentation, and the cirrus cloud pixel extraction result is obtained. The formula is as follows:

[0067]

[0068] In the Cumulus_mask image, pixels with a value of 1 represent cumulus clouds, and pixels with a value of 0 represent non-cumulus clouds, which is the result of sea ice pixel extraction.

[0069] Step 6: Estimation of sea ice concentration

[0070] This embodiment is based on Figure 7 The results of sea ice extraction from multi-band remote sensing images of HY-1C / D satellites are used to analyze the sea ice concentration SIC and sea ice coverage S ice The sea ice density is defined as the proportion of sea ice coverage per unit pixel area. It is mainly used in low and medium spatial resolution remote sensing images and is the basis for estimating sea ice coverage. The calculation is based on the 670nm red light band. The specific calculation steps are as follows:

[0071]

[0072] Where p is the pixel number, SIC p is the pixel sea ice density, ρ TOA，seawater is the apparent reflectance threshold of seawater pixel ε 5 , ρ TOA，ice is the apparent reflectivity value of the pure sea ice pixel. Among them, the apparent reflectivity of the sea water pixel ρ TOA，seawater The threshold is set to 0.05. When the apparent reflectance of a pixel is less than or equal to the threshold, it is considered to be a pure seawater pixel with a density of 0. The apparent reflectance of a pure sea ice pixel is ρ TOA，ice The threshold is obtained by counting the apparent reflectance value with the highest frequency in the sea ice pixels. When the pixel reflectance is greater than or equal to the threshold, it is considered to be a pure sea ice pixel with a density of 1; when the pixel apparent reflectance is between the two thresholds, the pixel is considered to be composed of a mixture of sea ice and sea water.

[0073] The calculation of sea ice density reflects the proportion of sea ice in the sea ice-seawater mixed pixel, and to a certain extent plays a role in demixing the mixed pixel. The sea ice coverage area calculated by combining the image pixel area is closer to the true value than the accumulation of the sea ice pixel area. The sea ice coverage area calculation formula is as follows:

[0074] S ice，p =A×SIC p

[0075] Where A is the image pixel area (1.21 km in this embodiment). 2 ), S ice，p is the sea ice coverage area in the pixel. The estimated results of sea ice density and sea ice coverage are as follows: Figure 8 As shown, the results show that the sea ice pixel area will overestimate the actual sea ice area to a certain extent.

[0076] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A sea ice remote sensing identification method based on ultraviolet reflectance difference, comprising the following steps: Step 1: Data preprocessing Preprocess multi-source remote sensing image data and convert optical band data into apparent reflectance ρ TOA Data, the optical band includes ultraviolet band, visible light band and near infrared band, and the thermal infrared band data is converted into brightness temperature data BT; Step 2: Identify the area of ​​interest For the image data preprocessed in step 1, use the vector file to crop out the area of ​​interest containing sea ice; Step 3: Spectral information enhancement For the apparent reflectance data ρ TOA Perform background water spectrum difference calculation, subtract background seawater spectrum pixel by pixel, and weaken the influence of background water color; normalize the background seawater difference data to obtain normalized difference reflectance data ρ' TOA ; Step 4: Feature band detection Select multiple typical spectral data of typical targets as training samples. The typical spectral data refers to the characteristic spectral data obtained by the spectrum determined as the designated target through the above steps 1-3. Take the average value of the characteristic spectral data of the training sample of each target, and calculate the contrast C between the spectral characteristic data according to the following formula: In the formula, ρ TOA,UV / RED is the apparent reflectance value of the typical target characteristic spectrum in the ultraviolet band or red light band, ρ TOA,VNIR is the apparent reflectance value of the characteristic spectrum of a typical target in the visible-near infrared band; the typical targets include sea ice, cirrus, cumulus and sea water; the two band combinations with the largest contrast C are selected for sea ice extraction and confusing target removal based on ultraviolet reflectance difference, and the final selected band combinations are: ultraviolet band and red light band combination, and ultraviolet band and near infrared band combination; Step 5: Identify sea ice based on the difference in ultraviolet band reflectance for the image. The specific steps are as follows: 1) For the apparent reflectance data ρ TOA The near infrared band grayscale image ρ TOA,NIR , use the threshold ε1 for threshold segmentation to obtain background seawater pixels and other pixels to be segmented. The specific formula is as follows: In the water_mask image, pixels with a value of 1 represent seawater, and pixels with a value of 0 represent non-seawater; 2) Use the Water_mask image as a mask to normalize the difference reflectance data ρ' TOA Operation is used to remove seawater pixels. For the normalized difference reflectance data ρ' of the removed seawater pixels TOA , use the ultraviolet band, green band, and red band to construct the corresponding index, and obtain the brightness temperature image data BT at the same time. Use the threshold ε2, threshold ε3, and threshold ε4 for threshold segmentation to obtain cirrus cloud pixels. The specific formula is as follows: Where, T1 = ρ' TOA,RED -ρ' TOA,GREEN , ρ' TOA,UV It is the normalized difference reflectance data of the ultraviolet band. The pixels with a value of 1 in the Cirrus_mask image represent cirrus clouds, and the pixels with a value of 0 represent non-cirrus clouds; 3) Use the Cirrus_mask image as a mask to normalize the difference reflectivity data ρ' TOA The operation is used to remove the seawater and cirrus pixels. For the normalized difference reflectance data ρ' after removing the seawater and cirrus pixels TOA , use the red light band and the near infrared band to construct the corresponding index, and obtain the brightness temperature image data BT at the same time. Use the threshold ε4 and the threshold ε5 for threshold segmentation to obtain cumulus pixels and sea ice pixels. The specific formula is as follows: Where T2 = ρ' TOA,NIR / ρ' TOA,RED , ρ' TOA,NIR is the normalized difference reflectance data in the near-infrared band, ρ' TOA,RED It is the normalized difference reflectance data of the red light band. The pixels with a value of 1 in the Cumulus_mask image represent cumulus clouds, and the pixels with a value of 0 represent non-cumulus clouds, which is the result of sea ice pixel extraction.

2. The method for sea ice remote sensing identification based on ultraviolet reflectance difference according to claim 1 is characterized in that It also includes step 6, sea ice concentration estimation, as follows: Based on the sea ice pixel extraction results in step 5, the most frequent apparent reflectance ρ is obtained by statistics. TOA The value is used as the apparent reflectance value of the pure sea ice pixel, combined with the apparent reflectance value of the sea water pixel, to calculate the sea ice concentration SIC; according to the sea ice pixel area, combined with the sea ice concentration, the sea ice coverage area S is calculated ice .

3. The method for sea ice remote sensing identification based on ultraviolet reflectance difference according to claim 1, characterized in that: In step 1, the multi-source remote sensing image includes an optical band of 355-865 nm, and thermal infrared bands of 1080 nm and 1200 nm, and the preprocessing includes geometric correction, radiation calibration, atmospheric correction and smoothing filtering.

4. The method for sea ice remote sensing identification based on ultraviolet reflectance difference according to claim 1, characterized in that: In the step 2, when dividing the area of ​​interest, the land is masked according to the map vector data.

5. The method for sea ice remote sensing identification based on ultraviolet reflectance difference according to claim 1, characterized in that: In step 3, the apparent reflectivity ρ is calculated using the following formula: TOA Subtract the background seawater value ρ TOA,seawater Then normalization is performed to achieve the effect of target enhancement: a i =ρ TOA,i -r TOA,seawater i=1,2,3,…,N r' TOA,i =(a i -a i,min ) / (a i,max -a i,min ) In the formula, i is the band number of the optical band, N is the number of optical bands of the remote sensing image, and a i is the background seawater difference of band i, a i,max and a i,min are the maximum and minimum values ​​in the difference spectrum curve of band i, respectively. Finally, the normalized difference reflectance data ρ' of band i is obtained. TOA,i .

6. The method for sea ice remote sensing identification based on ultraviolet reflectance difference according to claim 1, characterized in that: In step 5, threshold ε1=0.05, threshold ε2=0.4, threshold ε3=0.5, threshold ε4=266.5K, and threshold ε5=0.

75.

7. The method for sea ice remote sensing identification based on ultraviolet reflection difference according to claim 3 is characterized in that: In step 5 2) and 3), the brightness temperature image data BT uses the 1080nm thermal infrared band data.

8. The method for sea ice remote sensing identification and density estimation based on ultraviolet reflectance difference according to claim 1, characterized in that: In step 6, the sea ice concentration SIC and the sea ice coverage area S ice The calculation uses the red light band, and the formula is as follows: S ice,p =A×SIC p Where p is the pixel number, SIC p is the pixel sea ice density, ρ TOA,seawater ε5 is the apparent reflectance threshold of seawater pixel, ρ TOA,ice is the apparent reflectance value of the pure sea ice pixel, and the most frequent apparent reflectance ρ in the sea ice pixel is counted. TOA A is the image pixel area, S ice,p is the sea ice cover area in the pixel.

9. The method for sea ice remote sensing identification and density estimation based on ultraviolet reflectance difference according to claim 8, characterized in that: ε5=0.05.

Images