wri / raster2points

CLI to convert raster data to points (CSV or Pandas Dataframe)

raster2points

Convert one or multiple raster images to points. Tool will read first input raster and extract lat/lon coordinates and values for all pixels which have data. Optional it calculates geodesic area for each point based on pixel size. Successive input rasters will use data mask from first input raster.

Returns a Pandas dataframe, CLI will export results as CSV file.

Input files can be local file paths or S3 paths, or a mix. For reading from S3, you'll need AWS credentials configured, such as with a profile in ~/.aws and an AWS_PROFILE variable in your environment.

Multi-worker only works with S3 inputs, not local files.

Installation and Dependencies

This module uses rasterio and requires GDAL>=1.11. Use pip to install.

pip install raster2points

CLI Usage

raster2csv.py   [-h]
                [--col_names COL_NAMES [COL_NAMES ...]]
                [--separator {,,;,t}]
                [--max_block_size MAX_BLOCK_SIZE]
                [--calc_area [CALC_AREA]]
                [--workers WORKERS]
                INPUT [INPUT ...]
                OUTPUT

Python Usage

You can also use the module directly in python. It will return a Pandas dataframe with your data.

Get Pandas data frame

from raster2points import raster2df

raster1 = "path/to/file1.tif"
raster2 = "path/to/file2.tif"

df = raster2df(raster1, raster2, col_names=["name1", "name2"], calc_area=True)

print(df.columns)
print(df.dtypes)

df.head()

Export to TSV

from raster2points import raster2csv

raster1 = "path/to/file1.tif"
raster2 = "path/to/file2.tif"
output = "path/to/newfile.tsv"

raster2csv(raster1, raster2, output, col_names=["name1", "name2"], separator="\t", calc_area=True)

raster2points's Issues

Implement improved area calculations method

Is your feature request related to a problem? Please describe.
Improve performance for area calculations

Describe the solution you'd like
In gfw-raster-analysis-lambda you implemented an improved way of calculating the area.
Please implement here as well.

wri/gfw-raster-analysis-lambda@bf04302

add feature extract by mask