Satellite remote sensing method and device for detecting floating marine garbage

Abstract

The present invention discloses a satellite remote sensing method and device for detecting floating garbage in the ocean, including: obtaining the remote sensing reflectivity of the infrared band of the target area from the mixed spectrum of water garbage, and obtaining the remote sensing reflectivity of the infrared band of the water body at the same time, and calculating the first mixed spectrum index and the first water body spectrum index; obtaining the remote sensing reflectivity of the green light, red light and red edge bands of the target area from the mixed spectrum of water garbage, and calculating the second mixed spectrum index, obtaining the remote sensing reflectivity of the green light, red light and red edge bands of floating algae, and calculating the second water body spectrum index; determining the first spectrum threshold for distinguishing floating garbage from water body and the second spectrum threshold for distinguishing floating garbage from floating algae through the distribution of spectrum index, and identifying floating garbage according to the first spectrum threshold and the second spectrum threshold. The above technical scheme has strong practical operability, strong anti-interference and high recognition accuracy.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing technology, and in particular to a satellite remote sensing method and device for detecting marine floating garbage. Background Art

[0002] Marine garbage pollution is a major global environmental hotspot that has attracted great public attention. Marine garbage is mainly divided into floating garbage on the sea surface, beach garbage and seabed garbage. Among them, floating garbage is widely distributed and huge in quantity. Its main component is plastic garbage, accounting for as high as 60 to 80%, followed by wood products and paper products, accounting for 6.4% and 6.0% respectively. Therefore, most of the floating garbage in the sea has stable chemical properties and can exist on the sea surface for a long time, causing serious damage to marine life, ecosystems, food chains, etc.

[0003] Therefore, it is urgent to conduct scientific research on marine floating garbage. Among them, comprehensive and effective monitoring of marine floating garbage is the primary problem that needs to be solved. For example, where is the marine floating garbage distributed and how does its spatial distribution range change over time.

[0004] At present, the monitoring of the distribution of floating garbage in the ocean mainly relies on on-site observation. Although this method has a certain degree of recognition accuracy, it is time-consuming and labor-intensive, and the recognition range is very limited, making it difficult to be widely used. There are also theoretical studies on monitoring through satellite remote sensing, but it is still in its infancy. In addition, different types of garbage have different reflectance spectral characteristics, especially in the visible light band, which are easily disturbed during the recognition process. This brings certain difficulties to the remote sensing monitoring and identification of floating garbage in the ocean, resulting in low recognition accuracy. Summary of the invention

[0005] Purpose of the invention: The present invention provides a satellite remote sensing method and device for detecting floating marine debris, aiming to solve the technical problems of limited identification range and low identification accuracy of floating marine debris in the prior art.

[0006] Technical solution: The present invention provides a satellite remote sensing method for detecting floating garbage in the ocean, comprising: establishing a mixed spectrum of water body garbage; obtaining remote sensing reflectances of red edge and infrared bands of a target area from the mixed spectrum of water body garbage, and simultaneously obtaining remote sensing reflectances of red edge and infrared bands of a water body, and calculating a first mixed spectrum index and a first water body spectrum index; obtaining remote sensing reflectances of green light, red light and red edge bands of a target area from the mixed spectrum of water body garbage, and calculating a second mixed spectrum index; obtaining remote sensing reflectances of green light, red light and red edge bands of floating algae, and calculating a second water body spectrum index; determining a first spectrum threshold for distinguishing floating garbage from water body through the distribution of the first mixed spectrum index and the first water body spectrum index; determining a second spectrum threshold for distinguishing floating garbage from floating algae through the distribution of the second mixed spectrum index and the second water body spectrum index; identifying floating garbage according to the first spectrum threshold and the second spectrum threshold based on the spectrum data of the target area.

[0007] Specifically, the endmember spectra of floating garbage and water bodies are obtained to construct a mixed spectral dataset of water garbage.

[0008] Specifically, mixed spectral models are established for wood waste and plastic waste respectively, the mixed spectra of water waste under various floating waste coverage levels are calculated, and a water waste mixed spectral dataset is constructed.

[0009] Specifically, the mixed spectrum model is as follows:

[0010] R rs,t (λ)=(1-f)R rs Water (λ)+fR rs Debris (λ),

[0011] Among them, R rs,t (λ) represents the mixed spectrum of water garbage, f represents the coverage of floating garbage, R rs Water (λ) represents the end-member spectrum of water, R rs Debris (λ) represents the endmember spectrum of floating garbage.

[0012] Specifically, the average remote sensing reflectance of the red edge and infrared bands of the target area is calculated, and compared with the average remote sensing reflectance of the red edge and infrared bands of the measured target area without floating garbage water body, and the first mixed spectral index or the first water body spectral index is calculated.

[0013] Specifically, the following first spectral index model is adopted:

[0014] DSI 1 =R rsd (AVE)-R' rs w (AVE),

[0015] Among them, DSI 1 represents the first spectral index, R rs d (AVE) represents the average remote sensing reflectance of a specific red edge and infrared band in the mixed spectrum of water garbage, R' rs w (AVE) represents the average remote sensing reflectance of specific red edge and infrared bands of water bodies without floating garbage; the specific infrared bands include: red edge 2, red edge 3, near infrared and narrow wavelength near infrared.

[0016] Specifically, the average value of the remote sensing reflectance of the specific red edge and infrared band of the water body without floating garbage is calculated using the following formula:

[0017] R' rs w (AVE)=R rs rw (AVE)+nR rs wσ (AVE),

[0018] Among them, R' rs w (AVE) represents the average value of remote sensing reflectance of specific red edge and infrared bands of water body without floating garbage, R rs rw (AVE) represents the average value of remote sensing reflectance of specific red edge and infrared bands in multiple water areas without floating garbage, n represents the adjustment factor, R rs wσ (AVE) represents the standard deviation of the average value of remote sensing reflectance in a specific infrared band in multiple water areas without floating garbage.

[0019] Specifically, the mathematical red light band remote sensing reflectance is obtained by linear interpolation of the green light and red edge band remote sensing reflectances, and the ratio between the mathematical red light band remote sensing reflectance and the observed red light band remote sensing reflectance of the target area is calculated to obtain the second mixed spectral index or the second water body spectral index.

[0020] Specifically, the first mixed spectral index and the first water body spectral index under various floating garbage coverages are counted to determine the value that can maximize the distinction between floating garbage and water bodies, which is used as the first spectral threshold; the second mixed spectral index and the second water body spectral index under various floating garbage coverages are counted to determine the value that can maximize the distinction between floating garbage and floating algae, which is used as the second spectral threshold.

[0021] The present invention also provides a satellite remote sensing device for detecting floating garbage in the ocean, comprising: a spectrum establishment unit, a first spectrum index calculation unit, a second spectrum index calculation unit, a threshold calculation unit and a garbage identification unit, wherein: the spectrum establishment unit is used to establish a mixed spectrum of water garbage; the first spectrum index calculation unit is used to obtain the remote sensing reflectivity of the red edge and infrared band of the target area from the mixed spectrum of water garbage, and simultaneously obtain the remote sensing reflectivity of the red edge and infrared band of the water body, and calculate the first mixed spectrum index and the first water spectrum index; the second spectrum index calculation unit is used to obtain the target area from the mixed spectrum of water garbage The remote sensing reflectance of the green light, red light and red edge bands of the target area is calculated to obtain the second mixed spectral index; the remote sensing reflectance of the green light, red light and red edge bands of the floating algae is obtained to calculate the second water body spectral index; the threshold calculation unit is used to determine the first spectral threshold for distinguishing floating garbage from the water body through the distribution of the first mixed spectral index and the first water body spectral index; the second spectral threshold for distinguishing floating garbage from floating algae is determined through the distribution of the second mixed spectral index and the second water body spectral index; the garbage identification unit is used to identify the floating garbage according to the first spectral threshold and the second spectral threshold based on the spectral data of the target area.

[0022] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: strong practical operability, strong anti-interference ability and high recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of the flow chart of the satellite remote sensing method for detecting floating marine garbage provided by the present invention;

[0024] Figure 2 A histogram of the first spectral index provided by the present invention;

[0025] Figure 3 This is a histogram of the second spectral index provided by the present invention. DETAILED DESCRIPTION

[0026] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0027] The present invention targets four types of floating garbage commonly found in offshore areas (foam boards, buoys, plastic bottles and wood chips), and by deeply analyzing the reflectivity spectral characteristics of floating garbage, constructs a reflectivity spectral index that can highlight floating garbage signals and suppress background interference signals, and then establishes a remote sensing monitoring model applicable to various types of floating garbage. Therefore, the present invention has an optical theoretical basis, and the established model is widely applicable and has good generalizability. It also provides a technical reference for monitoring other types of marine floating garbage, and promotes the development of satellite remote sensing monitoring technology for marine floating garbage.

[0028] See also Figure 1 , which is a flow chart of the satellite remote sensing method for detecting floating marine garbage provided by the present invention.

[0029] In the embodiment of the present invention, a water garbage mixed spectrum is established.

[0030] In the embodiment of the present invention, endmember spectra of floating garbage and water bodies are obtained to construct a mixed spectrum data set of water body garbage.

[0031] In the embodiment of the present invention, mixed spectral models are established for wood garbage and plastic garbage respectively, mixed spectra of water garbage under various floating garbage coverages are calculated, and a mixed spectral data set of water garbage is constructed.

[0032] In the specific implementation, since the measured mixed spectrum data of water body garbage reflectance under different coverage of floating garbage is relatively limited, in order to deeply analyze the mixed spectrum characteristics of water body garbage reflectance under different coverage of floating garbage, and then more accurately determine the index threshold, the present invention adopts a simulation method to obtain a large amount of sufficient mixed spectrum data of water body garbage reflectance under different coverage of floating garbage, wherein the floating garbage is mainly foam board, buoy, plastic bottle and wood chip, and the simulation model adopted is a mixed spectrum model of the total radiation signal of the water surface considering the contribution of floating objects, which can be expressed as:

[0033] L t (λ)=(1-f)L w (λ)+fL p (λ), formula (1).

[0034] In formula (1), L t (λ) represents the mixed radiance of water and garbage, L w (λ) represents the water-leaving radiance without any floating objects, L p (λ) represents the radiance contributed by floating garbage, f represents the coverage of floating garbage, and the coverage of floating garbage can range from 1% to 100% with a step size of 1%.

[0035] In the specific implementation, both sides of formula (1) are divided by the downward solar irradiance E d (λ), the mixed spectrum model of water garbage can be expressed as:

[0036] R rs,t (λ)=(1-f)R rs Water (λ)+fR rs Debris (λ), formula (2).

[0037] In formula (2), Rrs,t (λ) represents the mixed spectrum of water garbage, f represents the coverage of floating garbage, R rs Water (λ) represents the end-member spectrum of water, R rs Debris (λ) represents the endmember spectrum of floating garbage.

[0038] In a specific implementation, the corresponding spectra can be calculated for various floating garbage coverages of each garbage type (foam boards, buoys, plastic bottles and wood chips), forming a water garbage mixed spectral dataset that includes a large amount of water garbage mixed spectral information.

[0039] In a specific implementation, the mixed spectra of remote sensing reflectance of water body garbage under different floating garbage coverage are different. Specifically, the mixed spectrum value of remote sensing reflectance of water body garbage will increase with the increase of floating garbage coverage.

[0040] In specific implementation, the present invention can simulate the mixed spectra of remote sensing reflectance of water body garbage under different floating garbage coverage (the range of floating garbage coverage is set to 1% to 100%, with a step size of 1%) for four types of floating garbage according to the actual coverage of floating garbage in the control experiment, and compare it with the mixed spectrum measured by the control experiment to verify the simulation accuracy of the mixed spectrum model, and then use it to study the response characteristics of the mixed spectrum of remote sensing reflectance of water body garbage as the floating garbage coverage increases. Furthermore, using the mixed spectrum model, the floating garbage coverage is increased by a certain step size, and a large number of mixed spectrum data of remote sensing reflectance of water body garbage under different floating garbage coverage are simulated to construct a mixed spectrum data set of water body garbage based on floating garbage coverage, which is used for the construction and verification of the subsequent floating garbage remote sensing detection and identification model. Finally, the above control experiment, field measurement and simulation of remote sensing reflectance spectrum data sets are resampled to the multi-spectral band of the target sensor according to the band setting of the target satellite sensor.

[0041] In an embodiment of the present invention, the remote sensing reflectance of the red edge and infrared band of the target area is obtained from the mixed spectrum of water garbage, and the remote sensing reflectance of the red edge and infrared band of the water body is obtained at the same time, and the first mixed spectrum index and the first water spectrum index are calculated.

[0042] In an embodiment of the present invention, the average remote sensing reflectance of the red edge and infrared band of the target area is calculated, and compared with the average remote sensing reflectance of the red edge and infrared band of the floating garbage-free water body in the measured target area to calculate the first mixed spectral index or the first water body spectral index.

[0043] In the specific implementation, based on the remote sensing reflectance end-member spectra of marine floating garbage measured by control experiments, and the remote sensing reflectance spectrum data of normal water bodies without floating garbage measured by offshore voyage surveys, the analysis revealed the unique spectral characteristics of floating garbage, and found that compared with the background water body, floating garbage has significantly higher reflectance spectral values, especially in the near infrared (NIR) and shortwave infrared (SWIR) bands. As for the four types of floating garbage, their reflectance spectra also show similarities and differences. The reflectance spectra of the four types of floating garbage in the visible light band are obviously different, while the differences in the NIR and SWIR bands are small. In particular, the reflectance of foam boards, polyvinyl chloride (PVC) buoys and wood chips all show similar high values ​​and spectral shapes. Based on the above spectral characteristics, in order to highlight the signals of the four types of floating garbage and suppress the signals of the background water body, the present invention uses the high reflection characteristics of the four types of floating garbage in the NIR band and the low reflection characteristics of the background water body to construct the first spectral index (Debris Spectral Index1, DSI) for detecting floating garbage. 1 ), the first spectral index can be constructed by comparing the average remote sensing reflectance of the red edge and infrared bands in the area with floating garbage with the average remote sensing reflectance of the infrared band in the area without floating garbage. The first spectral index model is as follows:

[0044] DSI 1 =R rs d (AVE)-R' rs w (AVE), formula (3).

[0045] In formula (3), DSI 1 represents the first spectral index, R rs d (AVE) represents the average remote sensing reflectance of a specific red edge and infrared band in the mixed spectrum of water garbage, R' rs w (AVE) represents the average remote sensing reflectance of specific red edge and infrared bands of water bodies without floating garbage; the specific red edge and infrared bands include: (Sentinel-2 satellite MSI sensor) Red Edge 2, Red Edge 3, Near Infrared (NIR) and Narrow Near-infrared (Narrow NIR).

[0046] In the embodiment of the present invention, the average value of the remote sensing reflectance of the specific red edge and infrared band of the water body without floating garbage is calculated using the following formula:

[0047] R' rs w (AVE) = R rs rw (AVE)+nR rs wσ (AVE), formula (4).

[0048] In formula (4), R' rs w (AVE) represents the average value of remote sensing reflectance of specific red edge and infrared bands of water body without floating garbage, R rs rw (AVE) represents the average value of the remote sensing reflectance of specific red edge and infrared bands in multiple water areas without floating garbage (in the target area), n represents the adjustment coefficient (can be set to 3, the specific value is determined according to the actual situation), R rs wσ (AVE) represents the standard deviation of the average value of remote sensing reflectance in specific red edge and infrared bands of multiple floating garbage-free water areas (within the target area).

[0049] In the specific implementation, R rs d The data information of (AVE) can come from the mixed spectrum data set of water garbage, R' rs w The data information of (AVE) can come from actual measurements.

[0050] In a specific implementation, the average remote sensing reflectance R of the above-mentioned red edge 2, red edge 3, near infrared (NIR) and narrow wavelength near infrared (Narrow NIR) is rs (AVE) can be calculated using the following formula:

[0051] R rs (AVE)=1 / 4(R rs (λ RE2 )+R rs (λ RE3 )+R rs (λ NIR )+R rs (λ NarrowNIR )), formula (5).

[0052] In formula (5), (R rs (λ RE2 ), R rs (λ RE3 ), R rs (λ NIR ) and R rs (λNarrowNIR ) represent the remote sensing reflectance of RedEdge 2, Red Edge 3, near infrared (NIR) and narrow wavelength near infrared (Narrow NIR), respectively.

[0053] In a specific implementation, the first mixed spectral index or the first water spectral index can be calculated using the above formula (3). When calculating the first mixed spectral index, R rs d The data of (AVE) refers to the mixed spectral data set of water garbage, and the mixed remote sensing reflectance of water garbage in specific bands under different floating garbage coverage for each type of garbage, R' rs w The (AVE) data refers to the actual measured remote sensing reflectance of a specific band without floating debris. Even in the absence of floating debris, there is still a difference between clear and turbid water. The remote sensing reflectance of multiple water areas in the target area can be obtained, and further effective reference can be made by adding the standard deviation of the average remote sensing reflectance of the water area.

[0054] When calculating the first water spectral index, similarly, R rs d The data of (AVE) still refers to the mixed spectral dataset of water garbage. The difference is that it refers to the remote sensing reflectance of water bodies in a specific band without floating garbage coverage (0 coverage, that is, pure water body without floating garbage), R' rs w The data reference of (AVE) is consistent with the calculation of the first hybrid spectral index, and is also the actual measured remote sensing reflectance of a specific band without floating garbage.

[0055] In an embodiment of the present invention, the remote sensing reflectance of the green light, red light and red edge bands of the target area are obtained from the mixed spectrum of water garbage, and the second mixed spectrum index is calculated; the remote sensing reflectance of the infrared band of floating algae is obtained, and the second water spectrum index is calculated.

[0056] In an embodiment of the present invention, a mathematical red light band remote sensing reflectance is obtained by linear interpolation calculation using green light and red edge band remote sensing reflectances, and the ratio between the mathematical red light band remote sensing reflectance and the observed red light band remote sensing reflectance of the target area is calculated to obtain a second mixed spectral index or a second water body spectral index.

[0057] In the specific implementation, in addition to the water background that needs to be eliminated during the floating garbage identification process, since large floating algae are also widely present on the ocean surface, such as the common Enteromorpha and Sargassum in the nearshore area, in order to effectively improve the anti-interference ability, a second spectral index is further introduced to distinguish floating algae.

[0058] In the specific implementation, the remote sensing reflectance multispectral data of large floating algae (Enteromorpha and Sargassum) obtained from the target satellite sensor (Sentinel-2 satellite MSI sensor) was used to analyze its reflectance spectral characteristics, and it was found that it also showed high reflectance characteristics in the near-infrared NIR band, which is very similar to the reflectance spectral characteristics of the four types of floating garbage studied in the present invention. This is bound to lead to confusion between marine floating garbage and large floating algae, and bring great errors to the remote sensing monitoring of marine floating garbage. Therefore, the present invention further constructs a second spectral index (Debris Spectral Index 2, DSI) to distinguish floating garbage from large floating algae. 2 ). Second spectral index DSI 2 The calculation formula is as follows:

[0059] DSI 2 =R rs '(λ RED ) / R rs (λ RED ), formula (6);

[0060] R rs '(λ RED )=R rs (λ GREEN )+((R rs (λ RE2 ) - R rs (λ GREEN )) / (λ RE2 - λ GREEN ))(λ RED - λ GREEN ), formula (7).

[0061] In formula (6) and formula (7), R rs '(λ RED ) represents the value at the red band obtained by linear interpolation of remote sensing reflectance of the green band and the red edge 2 band, R rs (λ RED ), R rs (λ GREEN ) and R rs (λ RE2 ) represent the remote sensing reflectance of the red light (near 665 nm), green light (near 559 nm) and red edge 2 (near 740 nm) bands of the Sentinel-2 satellite MSI sensor, respectively. RED , RE2 and λ GREEN They represent the red light (near 665 nm), red edge 2 (near 740 nm), and green light (near 559 nm) bands respectively.

[0062] In a specific implementation, the second mixed spectral index and the second water body spectral index can both be calculated using the above formula (6). When calculating the second mixed spectral index, R rs '(λ RED ) and R rs (λ RED ) is from the water garbage mixed spectrum dataset; when calculating the second water spectral index, R rs '(λ RED ) and R rs (λ RED ) data are referenced from actual measurements of the target area, obtained through observations by satellite sensors.

[0063] In an embodiment of the present invention, a first spectral threshold for distinguishing floating garbage from water bodies is determined by the distribution of a first mixed spectral index and a first water body spectral index; a second spectral threshold for distinguishing floating garbage from floating algae is determined by the distribution of a second mixed spectral index and a second water body spectral index.

[0064] In an embodiment of the present invention, a first mixed spectral index and a first water body spectral index under various floating garbage coverages are counted to determine a value that can maximize the distinction between floating garbage and water bodies, which is used as a first spectral threshold; a second mixed spectral index and a second water body spectral index under various floating garbage coverages are counted to determine a value that can maximize the distinction between floating garbage and floating algae, which is used as a second spectral threshold.

[0065] See also Figure 2 , which is the first spectral index histogram provided by the present invention.

[0066] In the specific implementation, in order to determine the threshold of the first spectral index for distinguishing floating garbage from water bodies, the present invention uses the established mixed spectral dataset of floating garbage and water body garbage under different coverage, and the remote sensing reflectance dataset of normal water bodies in the offshore without floating garbage obtained based on the field voyage survey data, and calculates the first spectral index DSI of floating garbage under different coverage respectively. 1 values ​​(generally ranging from -0.0111 to 0.2892), and the DSI of normal water bodies 1 The DSI values ​​under different floating garbage coverage were calculated. 1 Histogram of the distribution and DSI of each normal water sample 1 The ratio of floating garbage to non-floating garbage water in the overlapping part of the histogram of the floating garbage distribution is used to determine the first spectral index DSI that can best distinguish floating garbage from normal water. 1 Value, as the first spectral index DSI 1 The optimal threshold value of .

[0067] See also Figure 3 , which is a histogram of the second spectral index provided by the present invention. The units of the vertical axis on the left are the frequencies of four types of floating garbage, and the units of the vertical axis on the right are the frequencies of two types of large floating algae.

[0068] In a specific implementation, in order to determine the threshold of the second spectral index for distinguishing floating garbage from algae, the present invention uses the established mixed spectral dataset of floating garbage in water bodies under different coverage, and the multispectral data of remote sensing reflectance of large floating algae (mainly Enteromorpha and Sargassum) obtained from the target satellite sensor (Sentinel-2 satellite MSI sensor) image, and calculates the second spectral index DSI of floating garbage under different coverage. 2 values ​​(generally ranging from 0.9732 to 1.1056), and the DSI of large floating algae 2 The DSI values ​​under different floating garbage coverage were calculated. 2 Histogram of the distribution of macroalgae and DSI of each sample 2 The ratio of floating garbage to large floating algae in the overlapping part of the histogram of the floating garbage distribution is used to determine the second spectral index DSI that can best distinguish floating garbage from large floating algae. 2 Value, as the second spectral index DSI 2 The optimal threshold value of .

[0069] In a specific implementation, according to the above, the first spectrum threshold and the second spectrum threshold can be determined as 0 and 1.15 respectively.

[0070] In the embodiment of the present invention, floating garbage is identified according to the spectral data of the target area and the first spectral threshold and the second spectral threshold.

[0071] In a specific implementation, when the first spectral threshold and the second spectral threshold are determined to be 0 and 1.15 respectively, that is, when DSI 1 The value is greater than 0 and DSI 2 When the value is less than 1.15, the target pixel is detected as floating garbage; when DSI 1 When the value is less than 0, the target pixel is detected as background water. 1 The value is greater than 0 and DSI 2 When the value is greater than 1.15, the target pixel is detected as large floating algae. Using the validation data set and the confusion matrix method, the accuracy of the marine floating garbage detection and recognition established by the present invention is evaluated and analyzed. The results show that 92.75% of the floating garbage can be extracted and identified from the background water, and all the floating garbage can be accurately distinguished from large floating algae.

[0072] In the specific implementation, the remote sensing detection method of floating garbage in the ocean proposed in the present invention was applied to the Sentinel-2 satellite images before and after the passage of the super typhoon "Lekima" in 2019 near the coast of Wenling City, Zhejiang Province, and a case analysis of the application of remote sensing detection of floating garbage in the ocean was carried out, successfully detecting the distribution of floating garbage near the coast after the typhoon passed. Taking the distribution of floating garbage visually interpreted as the true value, the floating garbage extracted by the detection method of the present invention was compared and verified, and it was found that the model method proposed in the present invention has good practicality.

[0073] The present invention also provides a satellite remote sensing device for detecting floating garbage in the ocean, comprising: a spectrum establishment unit, a first spectrum index calculation unit, a second spectrum index calculation unit, a threshold calculation unit and a garbage identification unit, wherein: the spectrum establishment unit is used to establish a mixed spectrum of water garbage; the first spectrum index calculation unit is used to obtain the remote sensing reflectivity of the red edge and infrared band of the target area from the mixed spectrum of water garbage, and simultaneously obtain the remote sensing reflectivity of the red edge and infrared band of the water body, and calculate the first mixed spectrum index and the first water spectrum index; the second spectrum index calculation unit is used to obtain the target area from the mixed spectrum of water garbage The remote sensing reflectance of the green light, red light and red edge bands of the target area is calculated to obtain the second mixed spectral index; the remote sensing reflectance of the green light, red light and red edge bands of the floating algae is obtained to calculate the second water body spectral index; the threshold calculation unit is used to determine the first spectral threshold for distinguishing floating garbage from the water body through the distribution of the first mixed spectral index and the first water body spectral index; the second spectral threshold for distinguishing floating garbage from floating algae is determined through the distribution of the second mixed spectral index and the second water body spectral index; the garbage identification unit is used to identify the floating garbage according to the first spectral threshold and the second spectral threshold based on the spectral data of the target area.

[0074] The functions, steps and methods performed by the various execution units in the satellite remote sensing device for detecting floating marine debris provided by the present invention can refer to the functions, steps and methods in the satellite remote sensing method for detecting floating marine debris provided by the present invention.

Claims

1. A satellite remote sensing method for detecting floating marine debris, characterized in that: include: Obtain endmember spectra of floating garbage and water bodies, calculate mixed spectra of water garbage under various floating garbage coverage, and construct a mixed spectrum dataset of water garbage; From the mixed spectrum of garbage in the water body, the remote sensing reflectance of the red edge and infrared band of the target area is obtained, and the remote sensing reflectance of the red edge and infrared band of the water body is obtained at the same time, and the average remote sensing reflectance of the red edge and infrared band of the target area is calculated, and compared with the average remote sensing reflectance of the red edge and infrared band of the water body without floating garbage in the measured target area, and the first mixed spectrum index or the first water spectrum index is calculated; The mathematical red light band remote sensing reflectance is obtained by linear interpolation calculation of the green light and red edge band remote sensing reflectances of the target area, and the ratio between the mathematical red light band remote sensing reflectance and the observed red light band remote sensing reflectance of the target area is calculated to obtain a second mixed spectral index or a second water body spectral index, wherein the remote sensing reflectances of the green light, red light and red edge bands used to calculate the second mixed spectral index are from the mixed spectrum of water body garbage, and the remote sensing reflectances of the green light, red light and red edge bands used to calculate the second water body spectral index are from actual measurements of floating algae; Determine a first spectral threshold for distinguishing floating garbage from water bodies by the distribution of the first mixed spectral index and the first water body spectral index; determine a second spectral threshold for distinguishing floating garbage from floating algae by the distribution of the second mixed spectral index and the second water body spectral index; According to the spectral data of the target area, floating garbage is identified according to the first spectral threshold and the second spectral threshold.

2. The satellite remote sensing method for detecting floating marine garbage according to claim 1, characterized in that: The method of constructing a water garbage mixed spectral dataset comprises: Mixed spectral models were established for wood waste and plastic waste respectively, the mixed spectra of water waste under various floating waste coverages were calculated, and a water waste mixed spectral dataset was constructed.

3. The satellite remote sensing method for detecting floating marine debris according to claim 2, characterized in that: The mixed spectrum model is as follows: R rs,t (λ)=(1-f)R rs Water (λ)+fR rs Debris (l), Among them, R rs,t (λ) represents the mixed spectrum of water garbage, f represents the coverage of floating garbage, R rs Water (λ) represents the end-member spectrum of water, R rs Debris (λ) represents the endmember spectrum of floating garbage.

4. The satellite remote sensing method for detecting floating marine garbage according to claim 2, characterized in that: The calculation obtains the first mixed spectral index and the first water spectral index using the following first spectral index model: DSI1=R rs d (AVE)-R’ rs w (AVE), Where DSI1 represents the first spectral index, R rs d (AVE) represents the average remote sensing reflectance of the red edge and infrared bands in the mixed spectrum of water garbage, R' rs w (AVE) represents the average remote sensing reflectance of the red edge and infrared bands of the water body without floating garbage; the red edge and infrared bands include: red edge 2, red edge 3, near infrared and narrow wavelength near infrared.

5. The satellite remote sensing method for detecting floating marine debris according to claim 4, characterized in that: The average remote sensing reflectance of the red edge and infrared bands of the water body without floating garbage is calculated using the following formula: R’ rs w (AVE)=R rs rw (AVE)+nR rs wσ (AVE), Among them, R rs rw (AVE) represents the average value of the red edge and infrared band remote sensing reflectance of multiple water areas without floating garbage, n represents the adjustment coefficient, R rs wσ (AVE) represents the standard deviation of the average value of remote sensing reflectance in the red edge and infrared bands of multiple water areas without floating garbage.

6. The satellite remote sensing method for detecting floating marine debris according to claim 2, characterized in that: The determining of a first spectral threshold for distinguishing floating garbage from a water body, and the determining of a second spectral threshold for distinguishing floating garbage from floating algae, comprises: The first mixed spectral index and the first water body spectral index under various floating garbage coverages are counted to determine the value that can maximize the distinction between floating garbage and water bodies, which is used as the first spectral threshold. The second mixed spectral index and the second water body spectral index under various floating garbage coverages are counted to determine the value that can maximize the distinction between floating garbage and floating algae, which is used as the second spectral threshold.

7. A satellite remote sensing device for detecting floating marine debris, characterized in that: include: A spectrum establishment unit, a first spectrum index calculation unit, a second spectrum index calculation unit, a threshold calculation unit and a garbage identification unit, wherein: The spectrum building unit is used to obtain end-member spectra of floating garbage and water bodies, calculate mixed spectra of water body garbage under various floating garbage coverages, and construct a mixed spectrum data set of water body garbage; The first spectral index calculation unit is used to obtain the remote sensing reflectance of the red edge and infrared band of the target area from the mixed spectrum of water garbage, and simultaneously obtain the remote sensing reflectance of the red edge and infrared band of the water body, calculate the average value of the remote sensing reflectance of the red edge and infrared band of the target area, and compare it with the measured average value of the remote sensing reflectance of the red edge and infrared band of the water body without floating garbage in the target area, and calculate the first mixed spectral index or the first water body spectral index; The second spectral index calculation unit is used to obtain a mathematical red light band remote sensing reflectance by linear interpolation of the green light and red edge band remote sensing reflectances of the target area, calculate the ratio between the mathematical red light band remote sensing reflectance and the observed red light band remote sensing reflectance of the target area, and obtain a second mixed spectral index or a second water body spectral index, wherein the remote sensing reflectances of the green light, red light and red edge bands used to calculate the second mixed spectral index are from the mixed spectrum of water body garbage, and the remote sensing reflectances of the green light, red light and red edge bands used to calculate the second water body spectral index are from actual measurements of floating algae; The threshold calculation unit is used to determine a first spectral threshold for distinguishing floating garbage from water bodies through the distribution of the first mixed spectral index and the first water body spectral index; and to determine a second spectral threshold for distinguishing floating garbage from floating algae through the distribution of the second mixed spectral index and the second water body spectral index; The garbage identification unit is used to identify floating garbage according to the spectral data of the target area according to the first spectral threshold and the second spectral threshold.

Images