Imagery-based boundary identification for agricultural fields
Imagery-based boundary identification for agricultural fields is provided. In various embodiments, a time series of surface reflectance rasters for a geographic region is received. For each of the surface reflectance rasters, at least one index raster is determined, yielding at least one time series of index rasters. The at least one time series of index rasters is divided into a plurality of consecutive time windows. The at least one time series of index rasters is composited within each of the plurality of time windows, yielding a composite index raster for each of the at least one time series of index rasters in each of the plurality of time windows. The composite index rasters are segmented into a plurality of spatially compact regions of the geographic region. A plurality of polygons is generated from the plurality of spatially compact regions, each of the plurality of polygons corresponding to an agricultural field in the geographic region.
1. A method comprising:
receiving a time series of surface reflectance rasters for a geographic region;
determining, for each of the surface reflectance rasters, at least one index raster to produce one or more time series of index rasters;
dividing one or more time series of index rasters into a plurality of consecutive time windows;
compositing one or more time series of index rasters within each of the plurality of time windows, to produce a composite index raster for each of one or more time series of index rasters in each of the plurality of time windows;
segmenting the composite index rasters into a plurality of spatially compact regions of the geographic region by denoising the composite index raster using a spatial low pass filter; and
generating a plurality of polygons from the plurality of spatially compact regions, each of the plurality of polygons corresponding to an agricultural field in the geographic region.
2. The method of claim 1 , wherein the time series of surface reflectance rasters comprises satellite data.
3. The method of claim 1 , wherein the time series of surface reflectance rasters spans a growing season in the geographic region.
4. The method of claim 1 , wherein receiving the time series of surface reflectance rasters comprises determining surface reflectance from uncorrected reflectance data.
5. The method of claim 1 , wherein the at least one index raster comprises a normalized difference vegetation index raster.
6. The method of claim 1 , wherein the at least one index raster comprises a land surface water index raster.
7. The method of claim 1 , wherein the at least one index raster comprises a mean brightness raster.
8. The method of claim 1 , wherein determining the at least one index raster comprises downsampling the surface reflectance rasters.
9. The method of claim 1 , wherein the plurality of consecutive time windows correspond to early, mid-, and late phases of a growing season in the geographic region.
10. The method of claim 1 , wherein compositing comprises averaging the at least one time series of index rasters within each of the plurality of time windows.
11. The method of claim 1 , wherein segmenting comprises filling missing pixels in the composite index rasters.
12. The method of claim 11 , wherein filling missing pixels comprises applying linear interpolation to the composite index rasters.
13. The method of claim 1 , wherein segmenting comprises normalizing the composite index rasters.
14. The method of claim 1 , wherein segmenting comprises graph-based segmentation.
15. The method of claim 1 , wherein segmenting comprises Felzenszwalb segmentation.
16. The method of claim 1 , wherein generating the plurality of polygons comprises applying spatial smoothing to the plurality of spatially compact regions.
17. A system comprising:
a datastore comprising a time series of surface reflectance rasters for a geographic region; and
a computing node comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor of the computing node to cause the processor to perform a method comprising:
receiving the time series of surface reflectance rasters for a geographic region;
determining, for each of the surface reflectance rasters, at least one index raster to produce one or more time series of index rasters;
dividing one or more time series of index rasters into a plurality of consecutive time windows;
compositing one or more time series of index rasters within each of the plurality of time windows, to produce a composite index raster for each of one or more time series of index rasters in each of the plurality of time windows;
segmenting the composite index rasters into a plurality of spatially compact regions of the geographic region by denoising the composite index raster using a spatial low pass filter; and
generating a plurality of polygons from the plurality of spatially compact regions, each of the plurality of polygons corresponding to an agricultural field in the geographic region.
18. A computer program product for agricultural field boundary identification, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to perform a method comprising:
receiving the time series of surface reflectance rasters for a geographic region;
determining, for each of the surface reflectance rasters, at least one index raster to produce one or more time series of index rasters;
dividing one or more time series of index rasters into a plurality of consecutive time windows;
compositing one or more time series of index rasters within each of the plurality of time windows, to produce a composite index raster for each of one or more time series of index rasters in each of the plurality of time windows;
segmenting the composite index rasters into a plurality of spatially compact regions of the geographic region by denoising the composite index raster using a spatial low pass filter; and
generating a plurality of polygons from the plurality of spatially compact regions, each of the plurality of polygons corresponding to an agricultural field in the geographic region.