Shapely bbox geometry to sentinelhub bbox

Apps / Services Python SDK
1

I have list of polygons that are shapely type. These polygons are different boudning boxes that I have slected for certain process:

image

I want to have these polygons as bbox type. For that I have tried to use the get_bbox_from_shape function, but that doesn’t work as I understand it needs to have sentinelhub geometry:

def get_bbox_from_shape(shape,resolution):
    minx, miny, maxx, maxy = shape.geometry.total_bounds
    bbox_coords_wgs84=[minx, miny, maxx, maxy]
    bbox = BBox(bbox=bbox_coords_wgs84, crs=CRS.WGS84)
    bbox_size = bbox_to_dimensions(bbox, resolution=resolution)
    return bbox_size,bbox,bbox_coords_wgs84

get_bbox_from_shape(bbox_list[0],10)

AttributeError Traceback (most recent call last)
in
----> 1 get_bbox_from_shape(bbox_list[0],10)

in get_bbox_from_shape(shape, resolution)
1 def get_bbox_from_shape(shape,resolution):
----> 2 minx, miny, maxx, maxy = shape.geometry.total_bounds
3 bbox_coords_wgs84=[minx, miny, maxx, maxy]
4 bbox = BBox(bbox=bbox_coords_wgs84, crs=CRS.WGS84)
5 bbox_size = bbox_to_dimensions(bbox, resolution=resolution)

AttributeError: ‘Polygon’ object has no attribute ‘geometry’

So I have tried to convert my polygon into sentinelhub geoemtry in order to be able to get sentinelhub bbox type:

from shapely.geometry import shape, Polygon, MultiPolygon, MultiLineString
from sentinelhub import BBox, read_data, CRS, DataCollection,bbox_to_dimensions,geometry


for i in bbox_list:
    test=sentinelhub.geoemtry.Geoemtry(i,CRS.WGS84)

However, that return error:
NameError: name ‘sentinelhub’ is not defined

I’m not sure why I recieve this error as I have sentinelhub imported.
How can I convert my list f shapely bbox into a bbox that I can work with in sentinelhub?

maybe there is better/easier way to convert it?

**My end goal : to convert shapely bbox geometries into bbox I can work with with sentinelhub (class ‘sentinelhub.geometry.BBox’) **

2

There is a small error in your get_bbox_from_shape function: shape is already a shapely geometry, and as such, has no attribute geometry (this is what the first error says: AttributeError: ‘Polygon’ object has no attribute ‘geometry’)

Changing the line

minx, miny, maxx, maxy = shape.geometry.total_bounds

to

minx, miny, maxx, maxy = shape.bounds

works.

Beware that your code works correctly only for geometries that are in WGS84 coordinate reference system.

Your second approach is also possible, and has an error due to imports… If you do

from sentinelhub import BBox, CRS, Geometry

you can create bounding boxes from your geometries as

bbox = Geometry(shape, CRS.WGS84).bbox

(shape would be i in your for loop).

1 Like
3

Hi! @batic, I am getting this error, but can’t understand this, please help me out.
Thanks in anticipation

image
image1920×1080 211 KB

4

Hi @damsreserviors

The lower_left is not a method, but a property, so you can access it simply by calling bbox.lower_left (without the parenthesis).

NB: I am using sh-py version 3.8.0.

5

yes it worked, It is also mentioned in the documentation to use bbox.lower_left

6

@batic I am facing another problem, if you can help me please

image
image1920×1080 212 KB

7

I can’t say much from this; it seems like the water_mask variable (returned from get_water_mask_from_S2 method for this particular img[0,...,0], might not be well defined (or empty).

I suggest you try visualizing what is happening - e.g. plot the img[0,...,0], or debug the water_mask or similar.

8

Sorry @batic I forgot to send the SS, the values of BBox are following:

image
image1920×1080 220 KB

and the water mask is as follows:

image
image1920×1080 223 KB

9

Did you make it work then?

10

Yes, it did work, I had made some modifications in the water_mask function.