CN103778241A - Method for identifying large-scale vegetation degeneration area by remote sensing - Google Patents

Abstract

The invention discloses a method for identifying a large-scale vegetation degeneration area by remote sensing. The method comprises the following steps: 1) obtaining an MODIS time-serial vegetation index product data set MOD13Q1, and performing format and projection transformation; 2) performing splicing and clipping on the MODIS data, and synthesizing 16d NDVI data into a monthly NDVI value; 3) using an averaging method to obtain the average NDVI data year by year; 4) using the least square method to perform trend analysis on the independent variable time data and the dependent variable NDVI data; 5) using an Mann-Kendall nonparametric statistical test method to judge the significance of the trend; 6) performing overlay analysis on the results of trend analysis and the significance test results, so as to obtain NDVI changing trend data of a pixel scale, and dividing the results into a serious degradation area, a slight degradation area, a stable area, a slightly improved area and a significantly improved area; 7) counting a ratio of the space distribution to the area of each changing area based on the ArcGIS spatial analysis function, and performing driving factor analysis with the transportation network, the distribution of population, land use and other outside interferences at the same time.

Description

The recognition methods of a kind of large scale vegetation degeneration regional remote sensing

Technical field

The present invention relates to apply long-time sequence MODIS remotely-sensed data, based on the technical field in ENVI image processing software and ArcGIS spatial analysis software identification large scale vegetation degeneration region.

Background technology

Vegetation is as the main soil cover type in land, is the natural tie that connects atmosphere, soil and water body, in water and soil conservation, material and energy circulation and maintain the ecosystem and have a very important role aspect stablizing.And identify fast, accurately, objectively large scale vegetation degeneration region, can, for taking targetedly corresponding plans and measures that data supporting is provided, also can provide technical support for positive and negative feedback and Eco-Environmental Synthetic Analyses under climate change and artificial interference background simultaneously.Traditional research method is mostly on-the-spot investigation, and this method length consuming time, efficiency are low, is especially not suitable for the long-term vegetation Monitoring on Dynamic Change of big-and-middle yardstick, degenerates and follow the trail of and quantitative test.And current, dynamically become the key areas of global change research due based on remote sensing technology monitored area yardstick vegetation.Normalized differential vegetation index (NDVI) is one of actual parameter of assessment vegetation state, can react well vegetation growth state and vegetation coverage, is also the important information of weighing Regional Eco-environmental Quality.Meanwhile, wide coverage, temporal resolution are high and obtain free MODIS/NDVI product, for the monitoring of regional scale vegetation long-term dynamics provides reliable data source.

Summary of the invention

The present invention seeks to based on long-time sequence MODIS remotely-sensed data, process and trend analysis through professional software, thereby identify fast, accurately, objectively large scale vegetation degeneration region.

The present invention is take the MOD13Q1(NDVI product of 250m) as data source, obtain region average annual NDVI time series for many years by data pre-service, adopt least square method trend analysis and the Mann-Kendall method of inspection, judge vegetation variation trend feature, concrete steps are as follows:

Step 1 data are downloaded and pre-service

1) the MODIS time series vegetation index Product Data Set MOD13Q1(spatial resolution of download overlay area is 250m, temporal resolution is 16d), and utilize MRT software to carry out format conversion and (Hdf format conversion is Geotiff form, and Sinusoidal projection is converted to WGS84/Albers Equal Area Conic projection) changed in projection;

2) application ENVI software carries out mosaic splicing to MODIS data, and utilizes study area vector file to generate region of interest and carry out cutting, obtains time series MODIS/NDVI data set.16d NDVI data are synthesized a moon NDVI value by using formula (1) maximal value synthetic method (ENVI/band math) again, to eliminate the impact of the factors such as cloud, mist, sun altitude.

b=b ₁×（b ₁geb ₂）+b ₂×（b ₂gtb ₁） （1）

3) under ArcGIS platform, utilize averaging method to obtain by mean annual NDVI data, represent annual vegetation growth status, for covering trend analysis, vegetation provides data.

Step 2 trend analysis and significance test

4) to time independent variable and NDVI dependent variable data, adopt least square method, under ArcGIS, utilize formula (2) computational data to concentrate the NDVI of each pixel and the regression slope of time, formula is as follows:

$\mathrm{slope}=(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i{t}_i-\frac{1}{n}\left(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i\right)\left(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i\right))/(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t_i}^2-\frac{1}{n}{\left(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i\right)}^2)---\left(2\right)$

In formula: variable i is a year sequence number; x _iit is the NDVI value of i; t _ifor x _ithe corresponding time; N is studied length of time series.If the variation tendency of slope>0 explanation NDVI increases, otherwise reduces, be degeneration region.

5) Mann-Kendall is a kind of nonparametric statistics method of inspection, is used for judging the conspicuousness of trend, and computing formula is as follows: definition Z statistic is: $Z=\left\{\begin{array}{c}(S-1)/\sqrt{s\left(S\right)},S>0\\ 0,S=0\\ (S+1)/\sqrt{s\left(S\right)},S<0\end{array}\right.$ Wherein, $S=\underset{i=1}{\overset{n-1}{\mathrm{\&Sigma;}}}\underset{j=i+1}{\overset{n}{\mathrm{\&Sigma;}}}\mathrm{sgn}(x_j-x_i)$ $\mathrm{sgn}(x_j-x_i)=\left\{\begin{array}{c}1,x_j-x_i>0\\ 0,x_j-x_i=0\\ -1,x_j-x_i<0\end{array}\right.,$ s(S)=n(n-1)(2n+5)/18 (3)

In formula, sgn is-symbol function.Under given level of significance α, when | Z|> μ _1-α/2time, represent that sequence exists significant variation under alpha levels.Generally getting α is 0.05.

Step 3 analysis of trend and the identification of degeneration region

6) trend analysis result and Mann-Kendall significance test result are carried out to overlay analysis, obtain the NDVI variation tendency data on grid cell size, and result is divided into serious degradation district, slight degradation district, stable region, slight upgrading area and five types of remarkable upgrading areas.

7) based on ArcGIS spatial analysis functions, add up each variation partition space and distribute and area ratio, also the network of communication lines, population distribution and soil utilization are carried out to driving factors analysis as external interference simultaneously.

Advantage of the present invention:

The present invention can overcome tradition and take time and effort and have the deficiency of very large circumscribed ground investigation method, and based on long-time sequence MODIS/NDVI data, appliance computer image processing techniques, identifies large scale vegetation degeneration region fast, accurately, objectively.

Accompanying drawing explanation

Fig. 1 is the average annual NDVI space distribution of basin 2001-2013;

Fig. 2 is basin 2001-2013 vegetation variation tendency;

Fig. 3 is vegetation variation tendency Mann-Kendall significance test;

Fig. 4 is basin 2001-2013 vegetation variation tendency subregion;

Fig. 5 is vegetation variation tendency and geographic element spatial analysis.

Embodiment

Basin, Thousand-Island Lake is the important surface water seedbed of China, and its region vegetation growth status has great importance for water head site eco-environmental quality, identifies accurately, objectively vegetation degeneration region and can provide data supporting for water head site protection.Carry out instance analysis below in conjunction with this region:

Step 1 data are downloaded and pre-service

1) from NASA website, (http://reverb.echo.nasa.gov/reverb) downloads and covers the ranks number in basin and be respectively 28/5 and 28/6 MOD13Q1 data, time from calendar year 2001 to 2013 year totally 13 years 598 issue certificates.Utilize MRT software to carry out format conversion and projection conversion, Hdf format conversion is Geotiff form, and Sinusoidal projection is converted to WGS84/Albers Equal Area Conic projection.

2) application ENVI software carries out mosaic splicing to MODIS data, and utilizes study area vector file to generate region of interest and carry out cutting, obtains time series MODIS/NDVI data set.Use again maximal value synthetic method (ENVI/band math) that 16d NDVI data are synthesized to a moon NDVI value.

3) under ArcGIS platform, utilize averaging method to obtain 2001-2013 by mean annual NDVI data, represent annual vegetation growth status, for covering trend analysis, vegetation provides data.As Fig. 1, it is the average annual NDVI space distribution of basin, Thousand-Island Lake 2001-2013.

Step 2 trend analysis and significance test

4) to time independent variable and NDVI dependent variable data, adopt least square method, under ArcGIS platform, computational data is concentrated the NDVI of each pixel and the regression slope of time, as shown in Figure 2, on the occasion of representing that variation tendency increases, otherwise reduce, be degeneration region.

5) utilize the Mann-Kendall nonparametric statistics method of inspection to judge the conspicuousness of trend, significance test result is as Fig. 3.

Step 3 analysis of trend and the identification of degeneration region

6) regression slope region between-0.0005～0.0005 is divided into stable region, higher than 0.0005 be divided into and improve region, and lower than 0.0005 be divided into degeneration region.Mann-Kendall is checked the significance test result in 0.05 confidence level be divided into marked change (Z>1.96 or Z<-1.96) and change not significantly (1.96≤Z≤1.96) simultaneously.Finally, trend analysis result and Mann-Kendall assay are carried out to overlay analysis, obtain the NDVI variation tendency data on grid cell size, and result is divided into serious degradation district, slight degradation district, stable region, slight upgrading area and five types of remarkable upgrading areas, as shown in Figure 4.

7) based on ArcGIS spatial analysis functions, adding up each variation partition space distributes and area ratio, as shown in table 1, also can carry out driving factors analysis with external interference such as the network of communication lines, population distribution and soil utilizations simultaneously, disturb larger place as Fig. 5 shows to be subject to the mankind, vegetation degeneration is more for remarkable.

The shared area statistics analysis of the table 1 basin each subregion of vegetation variation tendency

In the description of this instructions, the description of reference term " embodiment ", " some embodiment ", " example ", " concrete example " or " some examples " etc. means to be contained at least one embodiment of the present invention or example in conjunction with specific features, structure, material or the feature of this embodiment or example description.In this manual, the schematic statement of above-mentioned term is not necessarily referred to identical embodiment or example.And specific features, structure, material or the feature of description can be with suitable mode combination in any one or more embodiment or example.

Although illustrated and described embodiments of the invention, those having ordinary skill in the art will appreciate that: in the situation that not departing from principle of the present invention and aim, can carry out multiple variation, modification, replacement and modification to these embodiment, scope of the present invention is limited by claim and equivalent thereof.

Claims (4)

1. large scale vegetation degeneration regional remote sensing recognition methods, it is characterized in that based on long-time sequence MODIS remotely-sensed data, obtain region average annual NDVI time series for many years by data pre-service, adopt least square method trend analysis and Mann-Kendall significance test method, judge vegetation variation trend feature, concrete steps are as follows:

MODIS time series vegetation index Product Data Set MOD13Q1 is obtained in step 1, download, and carry out format conversion and projection conversion through MRT software, be Geotiff form by Hdf format conversion, Sinusoidal projection is converted to WGS84/Albers Equal Area Conic projection;

Step 2, application ENVI software carry out mosaic splicing and cutting to MODIS data, then use maximal value synthetic method that 16 d NDVI data are synthesized to a moon NDVI value;

Step 3, under ArcGIS platform, utilize averaging method to obtain by mean annual NDVI data, for vegetation cover trend analysis data are provided;

Step 4, to time independent variable and NDVI dependent variable data, adopt least square method under ArcGIS, carry out trend analysis;

Step 5, utilize the Mann-Kendall nonparametric statistics method of inspection to judge the conspicuousness of trend;

Step 6, trend analysis result and significance test result are carried out to overlay analysis, obtain the NDVI variation tendency data on grid cell size;

Step 7, based on ArcGIS spatial analysis functions, add up each variation partition space and distribute and area ratio, also the network of communication lines, population distribution and soil utilization are carried out to driving factors analysis as external interference simultaneously.

2. a kind of large scale vegetation degeneration regional remote sensing according to claim 1 recognition methods, it is characterized in that, in described step 4, adopt least square method to carry out vegetation analysis of trend, if slope >0, the variation tendency that represents NDVI increases, otherwise reduce, be degeneration region.

3. a kind of large scale vegetation degeneration regional remote sensing according to claim 1 recognition methods, it is characterized in that, described step 5 utilizes conspicuousness that the Mann-Kendall nonparametric statistics method of inspection judges trend be changed significantly and change not significantly for being divided into according to the significance test result in 0.05 confidence level, wherein works as z>1.96 or zwhen <-1.96, represent to be changed significantly, when-1.96≤ z≤ 1.96, represent to change not remarkable.

4. a kind of large scale vegetation degeneration regional remote sensing according to claim 1 recognition methods, it is characterized in that, trend analysis result and significance test result are carried out overlay analysis by described step 6, and the result of described overlay analysis is divided into serious degradation district, slight degradation district, stable region, slight upgrading area and remarkable upgrading area.

Images