RU2497112C1 - Method for remote determination of degradation of soil cover - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: method involves probing an underlying surface having test areas with a multichannel spectrometer mounted on a space vehicle to obtain images on each channel; calculating, through zonal ratios of signal amplitude values in channels, partial degradation indices, specifically percentage content of humus (H), salinity index (NSI) and moisture loss index (W); determining the integral degradation index D based on a multi-parameter regressive relationship of the type: $$D={\left(\frac{H_0}{H}\right)}^{\mathrm{1,9}}\cdot {\left(\frac{NSI}{NS{I}_0}\right)}^{\mathrm{0,5}}\cdot {\left(\frac{W_0}{W}\right)}^{\mathrm{0,3}};$$

recalculating image brightness pixel values in the scale of the calculated degradation index for each pixel; selecting outlines of resultant images thereof with the established gradations of the degree of degradation. (H₀, NSI₀, W₀) are values of partial degradation indices for reference areas.

EFFECT: faster and more reliable determination of degree of degradation of soil cover.

5 dwg, 3 tbl

Description

The invention relates to agriculture and may find application in the compilation of the state inventory of arable land.

Soil is a multicomponent medium. By definition of the founder of soil science V.V. Dokuchaev [see N.F. Ganzhara, “Soil Science”, textbook, inter-regional association “Agro-education”, 2001, pp. 5-9] soil - an independent, natural-historical biocos natural body, which is an open four-phase dynamic system that includes a solid phase (polymer organic-mineral system), liquid phase (water, which occupies part of the feather space with dissolved organic and mineral substances), the gas phase (soil air), the living phase (soil biota, microorganisms inhabiting the soil).

As the industry develops, the territories adjacent to the industrial zones are subject to an ever-increasing flow of technogenic emissions and anthropogenic pressures leading to soil degradation.

However, at the moment there is no agreed list of both soil degradation indicators and methods for their determination. Maximum concentration limits for background levels of pollutants have not been established.

There is a method of assessing soil sustainability in arbitrary points (see, "Soil-ecological monitoring" textbook edited by DS Orlov, from Moscow State University, 1994, pp. 77-76, table 18, "Expert assessment of soil cover by indicators determining its integral stability (points) ”- analogue).

In the analogue method, 10 indicators were used to assess the integral stability of a certain group of soils: acidity, moisture, decomposition rate of plant residues, humus supply, relief, agricultural development, annual growth, base saturation, parent rocks, heat supply.

By the sum of the points, zones are identified: up to 14 points - unstable, from 18 to 20 - unstable, from 23 to 26 relatively stable and stable from 26 to 30 points.

The disadvantages of the analogue should be considered:

- subjectivity of a point expert assessment of each of 10 indicators;

- the inoperability and the complexity of the ground survey of large areas;

- unacceptably large number of private indicators that impede practical applicability;

- lack of documented boundaries of zones.

The scale of steps (from 0 to 6) of soil degradation is known: [see, magazine, “Geographical Bulletin”, No. 2 (17), 2011, article “Ecology and Nature Management”, pp. 49-50] - analogue.

Table 1 Criteria for determining the degree of soil degradation Criteria Degree of degradation 0 one 2 3 four 5 The area of the exposed humus horizon (A),% 0 <10 10-20 21-50 51-90 > 90 Power of abiotic sediment, cm 0 <2 2-10 11-20 21-40 > 40 Area of exposed parent rock (C) or underlying rock (D),% of total area 0-2 3-5 6-10 11-15 16-25 > 25 The decrease in the thickness of the soil profile (A + B),% of the original 0-1 1-3 3-25 26-50 51-75 > 75 The decrease in humus reserves in the soil profile (A + B),% of the original <5 5-10 11-20 21-40 41-80 > 80

table 2 Additional criteria for determining the degree of soil degradation Criteria Degree of degradation 0 one 2 3 four 5 The decrease in the content of trace elements (Mn, Co, Mo, B, Cu, Fe) in% of the average degree of security <5 5-10 11-20 21-40 41-80 > 80 The decrease in the content of mobile phosphorus, in% of the average degree of security <5 5-10 11-20 21-40 41-80 > 80 Decrease in degree of acidity (pH salt), in% of the average degree of acidity <5 5-10 11-15 16-20 21-25 > 25 Loss of soil mass t / ha / year <2 2-5 6-25 26-100 100-200 > 200

The disadvantages of the analogue method include:

- the uncertainty of the calculation of the integral indicator for different values of the private components;

- the subjectivity of determining the values of the private components, the inoperability and the great complexity of their quantitative calculation;

- lack of documentary measurements of the boundaries of the selected zones.

The closest analogue to the claimed technical solution is a method for remote measurements of the spectral brightness coefficient of soils and exposures using spectrometers (type SPI-2, SPI-74) mounted on an aircraft carrier [see L.I. Chapursky "Reflective properties of natural objects in the range 400-2500 nm", part 1, Ministry of Defense of the USSR, 1986, §5. Coefficients of Spectral Brightness (CSW) of Soils and Exposures, pp. 20-39]. In the closest analogue method, the QWs of various soil types are measured along the carrier flight path, which are a family of practically disjoint functions ρ (λ) whose ordinates increase with increasing wavelength (λ).

, (А=50; K=0,026; В=8,4) при коэффициенте корреляции между ρ₇₅₀ и Н_гум (%) порядка 0,9.Variations in the QWS, both in amplitude and in spectrum, depend on the concentration of humus, soil mineralization, and soil moisture. In particular, the empirical dependences of the QWN (ρ _λ = 750) at a wavelength of λ = 750 nm on the concentration of humus in the soil N _hum (%) of the form $${\rho }_{750}=A\times B\times \left(\mathrm{one}+\frac{\mathrm{one}}{K\times H_{g\mathrm{at}m}}\right)$$

, (A = 50; K = 0.026; B = 8.4) with a correlation coefficient between ρ ₇₅₀ and H _gum (%) of the order of 0.9.

The disadvantages of the method of the closest analogue are:

- the lack of an established integral relationship between soil degradation and measured values of the QWS;

- low spatial resolution of spectrometer type SPI, which does not allow us to highlight in the images the areas of degradation of soil cover with an area of less than ~ 2 km ² .

The problem solved by the invention consists in the operational, quantitative measurement of the degree of degradation of soil cover, by remote multispectral images, with the allocation and contouring of degradation sites, 0.1 hectares in size, suitable for compiling an inventory of arable land.

The technical result is achieved by the fact that the method for remote determination of soil degradation involves recording the brightness fields I (x, y) of arable land containing test reference plots with a multi-channel spectrometer mounted on an aerospace carrier in the zonal ranges 450-515, 525-605, 630-690 , 750-900 and 15 50-17 50 nm with the simultaneous receipt of digital images in each channel, the calculation of the partial state indices by a combination of the zonal relations of the signals I (x, y) in the channels for each pixel of the image; the index of humus content [H,%], the salinity index NSI and the moisture index NDWI, the choice as an integral criterion for soil degradation of the multi-parameter regression function of the product of the listed indices normalized relative to their values for the reference test plots in the form of power-law dependencies:

$$D={\left(\frac{H_0}{H}\right)}^{1.9}\cdot {\left(\frac{NSI}{NS{I}_0}\right)}^{0.5}\cdot {\left(\frac{NDW{I}_0}{NDWI}\right)}^{0.3}$$

recalculation by mathematical procedures of the values of each pixel of the image to the scale of the integral indicator D, the selection of the contours of degradation on the resulting matrix and their visualization in the form of processed images, where:

D - degree (category) of degradation 1, 2, 3, 4;

1.9; 0.5; 0,3 - indicators of sensitivity D to private indices.

The invention is illustrated by drawings, where:

figure 1 - spectral curves of humus horizons;

figure 2 - spectral curves of saline horizons (salt marshes);

figure 3 is a spectral characteristic of the soil cover in the band of 750 nm from the concentration of humus;

figure 4 - view of the processed image with contoured areas of degradation;

5 is a functional diagram of a device that implements the method.

The technical essence of the invention is as follows. A feature of remote sensing of the Earth from space is the dependence of the measurement signal on the characteristics of the environment (atmosphere, shooting conditions: angle of sight on the object, the height of the Sun, time of day, etc.), which necessitates the calibration of the measurement path. The latter in the claimed method is achieved by simultaneously shooting test (reference) sections, according to the values of the signals from which the calibration of the measurement paths of objects of the underlying surface is carried out in the future.

From mathematics it is known that for a unique solution of a system of equations, the number of equations must be equal to the number of unknowns.

Unknown in the claimed method are the indices - parameters: humus concentration, salinity, moisture loss. Therefore, for their unambiguous calculation, it is necessary to have at least three independent signals, which is realized by synchronous measurements in the indicated zones of the spectrometer.

In the quantitative calculation of the indices themselves, the parameters use various combinations of zonal relations [see, for example, Vinogradov B.V. "Aerospace Monitoring of Ecosystems", Science, M, 1984]. The main requirement for zonal relationships is the adequacy of measurements to the physical process, a high correlation coefficient between the signal and the calculated parameter, and resistance to external shooting conditions.

The concentration of humus in the soil (H) is calculated from the ratio [Vinogradov B.V.]:

, $$H[\%]=\frac{\mathrm{one}}{k}\cdot \ln \frac{{\rho }_0-{\rho }_{\min }}{{\rho }_{\lambda }-{\rho }_{\min }}$$

,

Where:

ρ _λ is the spectral reflection coefficient of the soil λ = 750-900 nm,

ρ ₀ - also for a humus-free parent rock,

ρ _min - also for multi-humus soil,

k - for soils of the middle band ~ 0.6.

, где $$SI=\sqrt{{\rho }_1^2+{\rho }_3^2}$$

,Salinity index $$NSI=\frac{S{I}_{\max }-SI}{S{I}_{\max }-S{I}_{\min }}$$

where $$SI=\sqrt{{\rho }_{\mathrm{one}}^2+{\mathrm{<figure-callout\; id="3"\; label="\rho "\; filenames="00000024.png,00000025.png"\; state="\{\{state\}\}">\rho </figure-callout>}}_3^2}$$

,

ρ ₁ , ρ ₃ - the QW of the pixel in the first and third channels.

To determine the normalized humidity index (NDWI), intensity measurements in the near infrared range ρ _NIR with a characteristic wavelength λ = 750-900 nm and in the short-wave infrared range ρ _SWIR with wavelengths λ = 1550-1750 nm are used. When calculating the NDWI index, the following formula is used:

$$NDWI=\frac{{\rho }_{NIR}-{\rho }_{SWIR}}{{\rho }_{NIR}+{\rho }_{SWIR}}$$

To identify hidden patterns between the measured indices - degradation parameters, and their contribution to the resulting (integral) degradation index, a multi-parameter regression function was used.

In accordance with the recommendations of the Mathematical Institute. Steklov, as monotone (non-discontinuous) functions, one should choose power functions of the form:

, $$D=a^x\cdot b^y\cdot c^z$$

,

Where:

D is an integral indicator of degradation;

a, b, c - indices-parameters that determine the degree of degradation;

x, y, z - indicators of the degree of sensitivity of a single index-parameter in the integral indicator.

The quantitative calculation of index-parameters and exponents of the multiparameter function is given in the example of a specific implementation.

An example implementation of the method

The claimed method can be implemented according to the scheme of figure 5. Functional diagram of the device contains the LandSat space system (1) on spacecraft (2) which has multichannel radiometers (3) that record the solar flux reflected from the underlying surface in the scanning strip (4), 185 km wide with a spatial resolution of ~ 30 m per pixel , in four zonal channels.

Route survey of planned regions is carried out according to commands from the onboard control complex (5) based on the daily program of spacecraft flight control transmitted from the system control center (6), at the request of consumers.

The measurement results are recorded in a buffer memory (7) and in the radio visibility zones of the spacecraft from ground-based reception points (8) (Moscow, Krasnogorsk, Novosibirsk) are dumped via the data transmission channel (9).

After preliminary processing of the information according to official features (number of revolution, time of shooting, coordinates), using the means (10) of the receiving point, the information goes to the data storage server (11).

Consumers perform thematic image processing at the processing center (12), where through the input and transmission device (13) information from the storage server enters an electronic computer (14) with a standard set of peripheral devices: a processor (15), random access memory (16) ), a magnetic disk drive (17), an information display device (18), a printing device (19), a keyboard (20).

The processed images of the plots with the calculated values of soil degradation for arable land in the region are placed on the storage server (21) of the processing results.

The condition of the soil cover was evaluated on the example of a site in the Ruzsky district of the Moscow region. 46 multispectral LandSat 4-5 TM images with adjacent surroundings were processed during the observation period from 2005 to 2011. The spatial resolution of the images is ~ 30 m / pixel, the square side of the treated area is 3030 m.

The condition of the non-degraded arable layer of loamy soils in the Moscow zone is estimated by the following indicators: thickness of the humus horizon 10-15 cm, humus reserves of about 50 t / ha, percentage of humus H ₀ (%) ~ 3%, salinity index ~ 0.03, moisture index ~ 0.24.

The averaged data of the signal processing results for the entire observation period are presented in table 3.

Table 3 Averaged processing results over the entire observation period Status Category Degree of degradation

Humus loss $$\frac{H_0}{H}$$

Salinity index Moisture loss $$\frac{W_0}{W}$$

one norm 1,0 1,0 1,0 2 low 1.4 1.12 1.05 3 average 1,62 1.3 1,07 four high 1.75 1,6 1.14

According to table 3, to obtain a multi-parameter regression function, write the following system of equations:

$$\{\begin{array}{l}D=2={\left(1.4\right)}^x\times {\left(1.12\right)}^y\times {\left(1.05\right)}^z\\ D=3={\left(\mathrm{1,62}\right)}^x\times {\left(1.3\right)}^y\times {\left(\mathrm{1,07}\right)}^z\\ D=\mathrm{four}={\left(1.75\right)}^x\times {\left(\mathrm{1,6}\right)}^y\times {\left(1.14\right)}^z\end{array}$$

After logarithm, the system of power equations reduces to linear algebraic. The solution of the system is carried out by the methods of Cramer and Sarrus. System identifier:

$$\Delta =\left|\begin{array}{ccc}\lg 1.4& \mathrm{<figure-callout\; id="1"\; label="lg"\; filenames="00000024.png,00000025.png"\; state="\{\{state\}\}"><figure-callout\; id="2"\; label="lg"\; filenames="00000024.png,00000025.png"\; state="\{\{state\}\}">lg</figure-callout></figure-callout>}\mathrm{1,2}& \lg 1.05\\ \mathrm{<figure-callout\; id="1"\; label="lg"\; filenames="00000024.png,00000025.png"\; state="\{\{state\}\}">lg</figure-callout>}\mathrm{1,62}& \lg 1.3& \mathrm{<figure-callout\; id="1"\; label="lg"\; filenames="00000024.png,00000025.png"\; state="\{\{state\}\}">lg</figure-callout>}\mathrm{1,07}\\ \lg 1.75& \mathrm{<figure-callout\; id="1"\; label="lg"\; filenames="00000024.png,00000025.png"\; state="\{\{state\}\}"><figure-callout\; id="6"\; label="lg"\; filenames="00000024.png"\; state="\{\{state\}\}">lg</figure-callout></figure-callout>}\mathrm{1,6}& \lg 1.14\end{array}\right|$$

Algebraic complements Δx, Δy, Δz are obtained by replacing the coefficients for unknowns with free terms, i.e. column from lg2, lg3, lg4. The following digital characteristics are obtained:

$$x=\frac{\Delta x}{\Delta }=1.9$$

$$y=\frac{\Delta y}{\Delta }=0.5$$

$$z=\frac{\Delta z}{\Delta }=0.3$$

For each pixel in the image, the indicator D is calculated and an integrated image matrix D (x, y) is formed.

Psychologically, the perception of the image by the human operator occurs at the level of the contour drawing. The outline drawing is obtained by calculating the gradient of the scalar function D (x, y) as:

$$gradD\left(x,y\right)=\frac{\partial D}{\partial x}\cdot i+\frac{\partial D}{\partial y}\cdot j$$

[cm. example "Derivative in direction" in the textbook N.S. Piskunov, “Differential and Integral Calculus for Higher Technical Schools,” 5th ed., Vol. 1, Science, M, 1964, pp. 264-268]

To obtain a contour drawing, a regular operator with an aperture of a window of 2 by 2 elements is selected.

i, j i, j + 1 i + 1, j i + 1, j + 1

Window elements are connected along diagonals (two mutually orthogonal directions) by a subtraction operation. The well-known poets of Sobel, Roberts, Laplace [see for example, Duda R.O., Hart P.E., "Pattern Recognition and Scene Analysis", trans. from English., M, Mir, 1976, pp. 287-288, §7.3 Spatial differentiation].

In particular, the Roberts operator R (i, j) at each point is calculated as:

$$R\left(i,j\right)=\left|D\left(i,j\right)-D\left(i+\mathrm{one,}j+\mathrm{one}\right)\right|-\left|D\left(i+\mathrm{one,}j\right)-D\left(i,j+\mathrm{one}\right)\right|$$

The points for which the value of R (i, j) is greater than the threshold value set by the operator are displayed.

The considered algorithm is implemented by the following program.

The program for selecting contours in the image

The results of determining the degradation of the soil cover of a test site in the Moscow region are shown in the form of degradation contours on a map diagram, illustrated by the figure in Fig. 4. Areas numbered (1) are water bodies, (2) are areas with low degradation, (3) are areas with medium degradation, (4) are areas with strong degradation.

The effectiveness of the claimed method is characterized by the efficiency, reliability, scale and documentation of the results.

Claims (1)

A method for remote determination of soil degradation, including the registration of brightness fields I (x, y) of arable land containing test reference sites, with a multi-channel spectrometer mounted on an aerospace carrier in the zonal ranges 450-515, 525-605, 630-690, 750-900 and 1550-1750 nm, with the simultaneous receipt of digital images in each channel, the calculation of private state indices by combining the zonal relationships of the I (x, y) signals in the channels for each pixel of the image, including the humus content index [H,%], index salting Nost NSI and humidity index NDWI, choice as integral criterion multiparameter regression function of soil degradation product of these indices, normalized with respect to their values for the reference test sites in the form of power relationships:
$$D={\left(\frac{H_0}{H}\right)}^{1.9}\cdot {\left(\frac{NSI}{NS{I}_0}\right)}^{0.5}\cdot {\left(\frac{NDW{I}_0}{NDWI}\right)}^{0.3},$$

recalculation by mathematical procedures of the values of each pixel in the image to the scale of the integral indicator D, the selection of the degradation contours on the resulting matrix and their visualization in the form of processed images, where
D - degree (category) of degradation 1, 2, 3, 4;
1.9; 0.5; 0,3 - indicators of sensitivity D to private indices.

Images