This project aims to produce a product of NDVI composed of several landsat scenes that can be separated in regions with their study periods considering the Brazilian semi-arid.
Note: It is possible to make modifications so that you can generate NDVI for areas outside the semi-arid territory.
Read this in other languages: English, Português
For build this project execute script build using following code.
sh build
To configure the project, you need to edit the samples/semi-arid/time_range.csv and samples/semi-arid/path_row.txt files. In time_range.csv, changing for each region its respective year and Julian day of the beginning and end of the study (rainy period + post rain) and the beginning of post-rain (for the algorithm to prioritize the best vegetation index pixels). For example, we have the following configuration:
REGION_NAME,START_DATE,END_DATE,START_POST_RAIN
MYREGION,2018 001,2018 091, 2018 070
In path_row.txt, you need to have the same regions added in time_range.csv, but now describing the path/rows of each landsat scene that makes up the region. For example, we have the following configuration:
MYREGION PATH01 ROW01 PATH02 ROW02 PATH03 ROW03 ...
Note: The configurations present in the code are correct to calculate the NDVI of the year 1986 of the Brazilian semi-arid.
To run the project use the following command:
python main.py run
python downloader.py download samples/semi-arid/path_row.txt samples/semi-arid/time_range.csv output/