Professional Documents
Culture Documents
Motivations
Council Regulation (EC) No 73/2009 Article 6 Good agricultural and environmental condition Minimum level of maintenance: Ensure a minimum level of maintenance and avoid the deterioration of habitats Retention of landscape features, including, where appropriate, hedges, ponds, ditches trees inline, in group or isolated and field margins Avoiding the encroachment of unwanted vegetation on agricultural land Protection of permanent pasture
Objective
Develop a methodology for identification of unwanted vegetation, abandoned lands from satellite images to be used in visual photointerpretation CAPI
Imput data
Satellite data : VHR Geoeye (26 may 2009) HR Spot 4 (9 May 2009) HR Spot 4 (14 Jun 2009) HR Spot 5 ( 14 Jul 2009)
Procedures
From satellite images vegetation indices NDVI , MSAVI , MMSAVI was generated .
NDVI was selected because is probably the most commonly used index in the last decades when studying vegetation. Its deficiencies and advantages have been thoroughly studied and are actually well known. This p rovided to the project results a kind of point of reference from where the performance of the other techniques was evaluated. NDVI is defined as: .
where R and NIR represent the surface reflectance on the red and near infrared regions of the spectrum, respectively.
MSAVI
The modified soil-adjusted vegetation index (MSAVI) and its later revision, MSAVI2, are soil adjusted vegetation indices that seek to address some of the limitation of NDVI when applied to areas with a high degree of exposed soil surface.
where RED is the red band reflectance from a sensor, NIR is the near infrared band reflectance, and L is the soil brightness correction factor. MSAVI uses the following formula to calculate L, where s is the slope of the soil line from a plot of red versus near infrared brightness values. :
Modified MSAVI
The formula of calculation is the same as in MSAVI , the only exception is instead of deducting RED band from NIR , we deducted GREEN band from NIR
After the stage where all 3 indices (NVDI,MSAVI, MMSAVI) Fig. 1 a stack was created , NDVI,MSAVI,MMSAVI become a layer of the new image obtained like this Fig. 2 , Fig. 3 . The stacks was classified using unsupervised classification ,obtained classe was gruped in 4 principal classes Fig. 5: Arable land Abandoned land Uncultivated land Uncultivated land encroachement of unwanted vegetation
NDVI 9 May
NDVI 14 Jun
NDVI 14 Jul
MSAVI 9 May
MSAVI 14 Jun
MSAVI 14 Jul
MMSAVI 9 May
MMSAVI 14 Jun
MMSAVI 14 Jul
Fig. 1
Stacks NDVI
MSAVI
MMSAVI
Fig. 2
Fig 3
Fig. 4
Fig. 5
Conclusions
Vegetation indices classification can be a great help in taking decisions regarding GAECs The method doesnt eliminate the classical visual interpretation on time series satellite images To do in the future: Supervised classification Initial masking of urban elements