If a processing request contains many polygons, the convex hull of those polygons which is submitted to CMR can sometimes be quite large and contain many granules that don't intersect any of the original polygons.
The client should implement an algorithm (k-means clustering) to create a set of smaller polygons that contain all of the original polygons and minimize the amount of space within each polygon that isn't also a part of one of the original polygons. The set should be no more than 5 to 10 polygons.
jpswinski
commented
2 years ago
If a processing request contains many polygons, the convex hull of those polygons which is submitted to CMR can sometimes be quite large and contain many granules that don't intersect any of the original polygons.
The client should implement an algorithm (k-means clustering) to create a set of smaller polygons that contain all of the original polygons and minimize the amount of space within each polygon that isn't also a part of one of the original polygons. The set should be no more than 5 to 10 polygons.