CN104181515A - Shallow sea water depth inversion method based on high-spectrum data of blue-yellow wave band - Google Patents

Abstract

The invention discloses a shallow sea water depth inversion method based on high-spectrum data of a blue-yellow wave band. Currently, models which perform depth inversion on clean water body through an optical remote sensing facility are mostly established for aiming at multiple-spectrum data. The algorithm is restricted by wide spectrum data wave band and little spectrum information. The inversion precision is largely affected by type of water body sediment. The invention provides a novel method for inversing shallow sea water depth of the clean water body through utilizing the high-spectrum data of the blue-yellow wave band (450-610 nanometers) according to a water body optical attenuation mechanism based on the high-spectrum data. The method can accurately extract shallow sea water depth distribution information within 30 meters. Furthermore one algorithm coefficient calibration is required for one remote sensor, thereby remarkably improving universality of the algorithm.

Description

A kind of shallow water depth inversion method based on blue-yellow wave band high-spectral data

Technical field

The present invention relates to a kind of application of satellitic remote sensing technology, particularly relate to the clean water body shallow water depth inversion technique of a kind of oceanic optical remote sensing.

Background technology

Solar radiation is subject to water body material molecule attenuation while propagating in water body mainly comprises absorption and scattering process, this effect can be by water surface visible light wave range spectral reflectivity different manifestations out, and along with the increase of water depth, damping capacity is more.For clean water body (a class water body), in water body, solar radiation is had to dissolved organic matter (CDOM), the Remote Sensing of Suspended Sediment Concentration of scattering process and there is the chlorophyll-a concentration of absorption low, solar radiation attenuation rate in water body is low, the transparent water depth of light increases, thereby can arrive sediment and arrive water surface through reflection, the variation of water surface emissivity spectrum also just can reflect the variation of water depth.

Based on above-mentioned physical process, experts and scholars have been developed a large amount of depth of water inversion algorithms both at home and abroad, and these algorithms comprise for the empirical algorithms of multispectral data feature (Clark et al, 1987; Lyzenga, 1981; Philpot, 1989) and half analytical algorithm (Lee et al., 2001; StumpfI, 2003, Adler-Golden et al., 2005, Albert and Gege, 2006), wherein empirical algorithms is difficult to realize general for different waters, different types of data, needs field measurement data to carry out model recurrence.And semi-analytical method also need to be corrected algorithm coefficient for different waters collection actual measurement bathymetric datas, and because multispectral data wide waveband, spectral information is limited, above-mentioned algorithm all supposes that sediment type is single, so its depth of water inversion error can increase along with the complexity of substrate type and significantly increase.

Along with the development of high spectrum resolution remote sensing technique, the particularly significantly raising of spaceborne high-spectrum remote sensing data signal to noise ratio (S/N ratio) and the increase of quantification gradation, for clean water body shallow water depth inverting provides new data source.High-spectral data can provide water body abundant spectral information, for accurately extracting shallow water depth, provides possibility.

Summary of the invention

In order to overcome above-mentioned the deficiencies in the prior art, the invention provides a kind of shallow water depth inversion method based on blue-yellow wave band high-spectral data.The method only need to be carried out an algorithm coefficient for high-spectral data load characteristic and be corrected the inverting that can be used for the clean water body shallow water depth of different substrate types.

The technical solution adopted in the present invention is: first high-spectral data is carried out to the data pre-service such as radiation calibration, atmospheric correction, solar flare are removed, white cap is corrected, land and water is cut apart, obtain seawater surface remote sensing reflectivity data.Then utilize field measurement bathymetric data to find the depth of water in Hyperspectral imaging close to the pixel of 0 (in 0.2 meter), the curve of spectrum that obtains its blue-yellow (450-610nm) wave band is as with reference to spectrum.Then calculate the spectrum angle between other pixel blue-yellow band spectrums and reference spectra within the scope of image waters, spectrum angle computing formula is shown in formula (3).Utilize actual measurement bathymetric data, by statistical regression, set up the funtcional relationship between relative water depth (relative reference pixel point) and spectrum angle, obtain depth of water inverting function.

Compared with prior art, the invention has the beneficial effects as follows, because spectrum angle can be compressed spectral information, Main Differences between outstanding spectrum, thereby reduced the impact of substrate type on inversion algorithm, in addition, the present invention is directed to a kind of remote sensor only need carry out the calibration of algorithm coefficient and can realize the depth of water inverting in different substrate types waters.

Accompanying drawing explanation

The shallow water depth inversion method flow process of Fig. 1 based on blue-yellow wave band high-spectral data

Embodiment

Below in conjunction with drawings and Examples, the specific embodiment of the invention is described in further detail.Following examples are used for illustrating the present invention, but are not used for limiting the scope of the invention.

Fig. 1 is for the present invention is based on blue-yellow wave band high-spectral data shallow water depth inversion method process flow diagram, and as shown in Figure 1, the clean water body shallow water depth flow process of inverting of the present invention is as follows:

(1) high-spectrum remote sensing data pre-service:

First high-spectrum remote sensing data to be processed is carried out to pre-service, comprise that proofread and correct on radiant correction, atmospheric correction, sea, extract in waters.First carry out radiant correction image DN value is converted into radiance value, computing formula is formula (1):

L _b＝Gain*DN _b+Bias (1)

Obtain after each wave band radiance image of high spectrum, adopt 6S atmospheric correction method or MODTRAN atmospheric correction algorithm to carry out atmospheric correction, obtain extra large surperficial remote sensing reflectivity data.The white cap effect causing due to the surperficial solar flare in sea and wind field can impact reflectivity data, therefore need to carry out to the sea surface remote sensing reflectivity data obtaining that solar flare is removed and white cap is corrected.To the remote sensing reflectivity data through above-mentioned processing, adopt improved water body index algorithm (MNDWI) to extract water body partial image.Water body index computing formula is formula (2)

MNDWI＝(Green-MIR)/(Green+MIR) (2)

In formula, Green represents green wave band, and MIR is middle-infrared band, and MNDWI is improved water body index, and MNDWI is greater than 0 pixel and is water body pixel.

(2) field measurement data set obtains:

Utilize multi-beam fathometer or other water depth detection instruments to obtain the actual measurement bathymetric data in test block, by global navigation satellite orientator, record the latitude and longitude coordinates of measurement point simultaneously, and by these locating information, these actual measurement bathymetric datas and high-spectrum remote sensing data are mapped, obtain field measurement data and high-spectrum remote-sensing reflectivity data corresponding data collection.

(3) calculate at spectrum angle:

By the actual measurement bathymetric data of having good positioning, in Hyperspectral imaging, search out the depth of water close to the pixel of 0 (in 0.2 meter), obtain the curve of spectrum in its 450～610nm wavelength band, take this curve of spectrum as reference spectra, 450nm～610nm curve of spectrum that in calculating image, each water body pixel is corresponding and the spectrum angle between this reference spectra curve.Spectrum angle computing formula is shown in formula (3).

$\mathrm{\θ}=\arccos \frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i\·y_i}{\sqrt{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i^2}\sqrt{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{y}_i^2}},\mathrm{\θ}\∈[0,\frac{\mathrm{\π}}{2}]---\left(3\right)$

In formula, x _ifor with reference to reflectance spectrum, y _ifor reflectance spectrum corresponding to image pixel, n is the wave band number between 450nm～610nm.

(4) model coefficient is corrected

Utilize each pixel reflectance spectrum that part field measurement bathymetric data and reference point depth of water difference and previous calculations obtain and the spectrum angle between reference spectra to carry out least square regression matching, through test, both are exponential function relation, and relational expression is as follows:

Z-Z ₀＝e ^(ax+b)

In formula, Z represents pixel corresponding depth, Z ₀represent the reference point depth of water (close to 0), x is the spectrum angle of each pixel relative reference spectrum, a and the b coefficient for needing to demarcate.By measured data, undertaken obtaining the depth of water inversion algorithm for this spectroscopic data sensor characteristics after function coefficients calibration.

(5) depth of water inversion accuracy checking:

Utilize through the depth of water inversion algorithm of coefficient calibration high-spectrum remote-sensing reflectivity image water body region is processed, obtain shallow water depth distributed data, utilize other depth of water measured datas to carry out precision test to arithmetic result, calculate root-mean-square error and mean deviation.

Claims (6)

1. the shallow water depth inversion method based on blue-yellow wave band high-spectral data, is characterized in that, the described inversion method based on high-spectral data spectral signature comprises the following steps:

S1: high-spectral data is carried out to the remote sensing reflectivity data that data pre-service obtains water body region.

S2: on-the-spot depth of water measured data gathers and location;

S3: utilize field measurement data acquisition to approach 0 meter of (in 0.2 meter) depth of water point 450nm～610nm Hyperspectral imaging remote sensing reflectance spectrum most as with reference to spectrum, calculate the spectrum angle between the corresponding spectrum of other water body pixels and reference spectra.

S4: utilize part measured data, with the spectrum angle calculating, depth of water inverting exponential function is carried out to coefficient and correct, obtain the depth of water inversion algorithm for this high spectrum sensor feature.

S5: utilize the aforementioned depth of water inversion algorithm obtaining to carry out the inverting of the high-spectral data depth of water, and utilize field measurement data to carry out precision test.

2. the shallow water depth inversion method based on blue-yellow wave band high-spectral data according to claim 1, is characterized in that, the high-spectrum remote sensing data pre-service in described S1 comprises that proofread and correct on radiant correction, atmospheric correction, sea, extract in waters.

3. the shallow water depth inversion method based on blue-yellow wave band high-spectral data according to claim 1, is characterized in that, the on-the-spot depth of water measured data collection in described S2 and location comprise that boat measurement bathymetric data gathers and corresponding GPS position data collecting.

4. the shallow water depth inversion method based on blue-yellow wave band high-spectral data according to claim 1, is characterized in that, described S3 comprises:

S31: utilize actual measurement bathymetric data and GPS locating information to find close to 450nm～610nm spectroscopic data of 0 meter of depth of water as with reference to spectrum;

S32: the spectrum angle in calculating remote sensing image between the corresponding 450nm～610nm spectroscopic data of each pixel and reference spectra, spectrum angle computing formula is referring to formula (3).

5. the shallow water depth inversion method based on blue-yellow wave band high-spectral data according to claim 1, is characterized in that, the relative reference depth of water point depth of water and the funtcional relationship between spectrum angle in described S4 are exponential function relation, that is: Z-Z ₀=e ^(ax+b), in formula, Z represents pixel corresponding depth, Z ₀represent the reference point depth of water (close to 0), x is the spectrum angle of each pixel relative reference spectrum, a and the b coefficient for needing to demarcate.In funtcional relationship, coefficients by using least square fitting method is demarcated.

6. the shallow water depth inversion method based on blue-yellow wave band high-spectral data according to claim 1, is characterized in that, the depth of water inversion accuracy checking in described S5 adopts the actual measurement depth of water and inverting depth of water root-mean-square error and mean deviation to carry out accuracy evaluation.