HJ 1008-2018Technical specification for straw burning monitoring based on satellite remote sensing (English PDF)

4.1 Monitoring principle

Differences in brightness temperature between mid-infrared and thermal infrared based on suspected fire point pixels and background normal temperature pixels in straw burning Do not heat abnormal points, combined with land classification data, extract suspected fire points from straw burning.

4.2 Monitoring means

Straw incineration monitoring is carried out using optical and infrared sensors of polar-orbiting satellites. The sensor should have a range of 0.65µm 2 visible light red band, near-infrared band near 0.8µm, mid-infrared band near 4µm, near 11µm and 12µm refer to Appendix A for the thermal infrared band and band settings.

4.3 Monitoring content

The location, quantity, and reliability of suspected fire spots in straw burning.

4.4 Monitoring process

The satellite remote sensing monitoring process for straw burning is shown in Figure 1.

Satellite remote sensing data Remote sensing data preprocessing Straw burning suspected fire point extraction Estimation of reliability of suspected fire point in straw burning Land use data Monitoring products to produce geographic information data Thermal anomaly extraction

Figure 1 Straw burning satellite remote sensing monitoring process

5.1 Remote sensing data preprocessing

First, the satellite remote sensing data is subjected to quality inspection, radiation correction and geometric correction, and then the visible light and near infrared waves are calculated.

The apparent reflectance of the segment and the brightness temperature of the mid-infrared and thermal infrared bands.

The apparent reflectance is calculated as follows.

Cos

ESUN

In the formula.

rho -- apparent reflectance;

--constant, sr;

- radiance, W/(m2 · sr · µm);

-- the distance between the sun and the earth, the astronomical unit;

ESUN - the average solar spectrum irradiance at the top of the atmosphere, W/(m2 · µm);

theta -- solar zenith angle, °.

The brightness temperature is calculated as follows.

Ln 1

In the formula.

-- apparent brightness temperature, K;

-- speed of light, m/s;

Lambda--center wavelength, µm;

- radiance, W/(m2 · sr · µm);

h--Planck constant, taking 6.626×10-34J·s;

k - Boltzmann constant, taking 1.38 × 10-23 J/K.

5.2 Cloud and water body pixel recognition and rejection

Identify and eliminate cloud and water pixels that meet the following conditions.

The discriminating conditions of the cloud pixel are.

rho1 12 t1 rho2 12 t 2 12 t1 Rn( )rn Th T Th Th T Th T Th rho rho rho rho ) > or ( and and delta > and and delta > Daytime at night (5) in the formula.

4T - the brightness temperature of the pixel in the mid-infrared band (around 4µm), K;

Tdelta - the difference between the brightness temperature of the pixel in the mid-infrared band (around 4 µm) and the thermal infrared band (near 11 µm), K;

rho4Th

- discriminate the threshold, the reference value may take 0.3;

t3Th -- discriminating threshold, the reference value can be taken as 300K;

t4Th -- discriminating threshold, the reference value can be 305K;

t1Thdelta -- The threshold is determined, and the reference value can be taken as 10K.

5.3.3 Initial fixed fire point absolute threshold test

If the initial fire point is in the daytime, the formula (6) is satisfied and it can be judged as a tentative fire point. In the nighttime situation, satisfied Equation (7) can be identified as a thermal anomaly point. Other initial fire points that do not meet the conditions need to enter the background threshold test process.

Determine it in one step.

4 t5T Th > (6)

4 t6T Th > (7)

In the formula.

t5Th -- discriminating threshold, the reference value can be taken as 360K;

t6Th -- discriminates the threshold, and the reference value can be 320K.

5.3.4 Initial fixed fire point background threshold test

Centering on the initial fire point, create a background window of size N×N, classify and count the background pixels in the window.

5 its brightness temperature characteristics. Background pixels include both background fire point pixels and valid background pixels.

Among them, the background fire point pixels meet the following conditions during daytime and nighttime respectively.

. 4 t7 t2( ) ( )T Th T Thdelta > and delta > (8) Night. 4 t8 t3( ) ( )T Th T Thdelta > and delta > (9) in the formula.

t7Th -- discriminating threshold, the reference value can be 325K;

t2Thdelta - discriminating threshold, the reference value can be taken as 20K;

t8Th -- discriminating threshold, the reference value can be taken as 310K;

t3Thdelta -- The threshold value can be taken as 10K.

The cloudless land background pixels outside the background fire point pixel in the window are valid background pixels. If the number of valid background pixels Satisfy 25% of the total number of cells in the window, and more than 8, then the background pixel temperature characteristics of the statistics window, the window start size it is 3×3. If the effective background pixels are not enough, increase the window (eg. 5×5, 7×721×21) and continue the above Classify and count until there are enough valid background pixels in the window. If N=21 is still not enough effective background for the pixel, the initial fire point is identified as a non-fire point.

If the above background fire point pixel and effective background pixel temperature characteristics are successfully extracted, then the temperature of the initial fire point is the characteristics (4T, 11T, and Tdelta) are used to determine multiple threshold conditions as follows.

E1 TT T Th deltadeltadelta > delta × (10) t4T T Thdeltadelta > delta (11) 4 4 e2 4T T Th delta > × (12) 11 11 11 t9TT Thdelta > - (13) 4 t10Thdelta ' > (14) in the formula.

Tdelta -- the brightness of the effective background pixels in the mid-infrared (near 4 µm) and thermal infrared (near 11 µm) the mean value of the difference, K;

- the mean value of the luminance temperature of the effective background pixel in the mid-infrared band (around 4 µm), K;

11T - the brightness temperature of the pixel in the thermal infrared band (near 11µm), K;

11T -- the average of the luminance temperature of the effective background pixel in the thermal infrared band (near 11 µm), K;

4delta - the average absolute deviation of the luminance temperature of the effective background pixel in the mid-infrared band (around 4 µm);

11delta - the average absolute deviation of the luminance temperature of the effective background pixel in the thermal infrared band (near 11 µm);

delta delta -- Brightness of effective background pixels in the mid-infrared (near 4 µm) and thermal infrared (near 11 µm) the average absolute deviation of the temperature difference;

4delta ' -- the average absolute deviation of the brightness temperature of the background fire point pixel in the mid-infrared band (around 4 µm);

e1Th - discriminate the threshold, the reference value is taken as 3.5;

t4Thdelta - discriminating threshold, the reference value is taken as 6K;

e2Th -- discriminates the threshold, and takes a reference value of 3;

t 9Th -- discriminating threshold, the reference value is taken as 4K;

t10Th -- The threshold is determined, and the reference value is taken as 5K.

If the initial set fire point satisfies all the conditions in equations (10)~(12), it also satisfies formula (13) or (14).

In one of the six conditions, the initial fixed fire point is identified as a tentative fire point; all conditions in equations (10) to (12) are met at night.

At the time, the initial fire point is identified as a hot abnormal point, otherwise it is identified as a non-fire point.

The daytime is identified as a tentative fire point for the false fire point removal to obtain a thermal anomaly point.

5.3.5 False fire point removal

5.3.5.1 False fire point removal caused by solar flare

Calculate the flare angle of the fire point pixel.

Cos cos cos sin sin cos Gvsvs theta theta theta theta theta phi= - (15) in the formula.

-- the flaming angle of the fire point pixel, °;

-- Observing the zenith angle, °;

--Sun zenith angle, °;

phi -- relative azimuth, °.

The threshold conditions for determining solar flare are.

1g aThtheta and > (17) 3( ) 0g a wTh Ntheta ( ) (18) in the formula.

1aTh -- discriminate the threshold, the reference value is 2 °;

2aTh -- discriminate the threshold, the reference value is 8 °;

3aTh -- discriminant threshold, the reference value is taken as 1 2 ° ;

rho5Th

- discriminate the threshold, the reference value is taken as 0.1;

rho6Th

- discriminate the threshold, the reference value is taken as 0.2;

N -- The number of water body pixels in the statistics window.

If the tentative fire point pixel satisfies one of the conditions in equations (16) to (18), it is determined to be false due to solar flare.

Fire point.

5.3.5.2 False fire point removal at the edge of the desert

Set threshold conditions for identifying false fire points for the radiation characteristics of the desert edge.

rho7

4 t11

4 t12

4 4 e4 4

N Th N

N Th

T Th

TT Th

『 6 Monitoring product production Monitoring products include daily, monthly, quarterly, annual reports, etc. The report should include text, thematic maps and statistical forms. Text the word information is information describing the results of satellite remote sensing straw burning, including time, range, satellite and sensor, monitoring fire the distribution, number, etc. of points. The special map of straw burning includes map name, legend, scale, and suspected fire point distribution information of straw burning and administrative area geographic information, traffic roads and airport information. The statistical table includes the longitude, latitude, and fire point of the fire.

Information such as the name and statistics of the political district.

The statistics of the number of fire points in the daily straw burning monitoring products are based on the statistics of each pixel of the single-day product of the day.

If two or more sensors monitor the same location fire point in a day, do not accumulate counts for a specific administrative area.

The statistics of the number of internal fire points are the sum of the values of the fire points occurring in the administrative area on that day. Monthly, seasonal and annual straw burning monitoring production the statistics of the number of fire points in each administrative area of the product are the values of the fire spot pixels in the products of the current month, the current season and the current day.

Sum.

7.1 Satellite data quality

Before data preprocessing for suspected fire point information in straw burning, it is necessary to ensure the quality of remote sensing raw data and avoid Data with equal quality issues participate in subsequent processing, resulting in misidentification of the results.

7.2 Geometric positioning accuracy

Before using the remote sensing data of different satellites and sensors for fire point monitoring, the geometric position registration and registration should be guaranteed.

Degree is within a pixel.