Issues in PlanetScope Images in Egypt

Apps / Services
1

Hi everyone,

We download daily images from PlanetScope for a region in south Egypt. We faced the following issues in the past days:

  • Until yesterday, there was no images available in March for this region.

  • Today, we can see that there are only 3 images available (one from March 4, and two from March 5). We would expect to find at least one image available everyday.

  • Two of the 3 images available in March have properties missing in their metadata, like “visible_percent” and “clear_percent”, among other properties that we use in our application.

Can you help us solve these issues?

Below is the coordinates of a farm from that region. Let us know if you need other information.

{“type”: “Polygon”,
“coordinates”: [[[28.85965678475845, 22.383747601558337],
[28.85965678475845, 22.39145000516439],
[28.867739956468856, 22.39145000516439],
[28.867739956468856, 22.383747601558337],
[28.85965678475845, 22.383747601558337]]]}

2

Hi Karim,

Planet is promising almost daily imagery. It happens that in some days observation is not possible due to weather or occasional technical glitch.
In your AOI I can also data from 2nd, 4th, 5th and 8th of March:

Screenshot 2022-03-09 at 13.02.57
Screenshot 2022-03-09 at 13.02.572880×1472 278 KB

What do you mean that the properties are missing in the metadata? Where specifically are you looking at?

Best,
Grega

3

I am using the following function to search for the available images

def NativeSearch(oauth, geo, startdate, enddate, maxcloudpercent):
url = “https://services.sentinel-hub.com/api/v1/dataimport/nativesearch”
payload = {
“provider”: “PLANET”,
“item_types”: [“PSScene4Band”],
“filter”: {
“type”: “AndFilter”,
“config”: [
{“type”: “GeometryFilter”, “field_name”: “geometry”, “config”: geo},
{
“type”: “DateRangeFilter”, “field_name”: “acquired”,
“config”: {“gte”: isodate(startdate), “lte”: isodate(enddate, is_end=True)}
},
{
“type”: “RangeFilter”, “field_name”: “cloud_cover”, “config”: {“lte”: maxcloudpercent/100} # We use maxcloudpercent=100
},
{“type”: “StringInFilter”, “field_name”: “quality_category”, “config”: [“standard”]},
{“type” : “PermissionFilter”, “config” : [ “assets:download” ]},
{“type” : “AssetFilter”, “config” : [ “analytic_sr” ]}
]
}
}
response = oauth.post(url, json=payload)
response.raise_for_status()
return response

4

I run our default native search with the same polygon and I got 7 results:

"acquired": "2022-03-08T08:33:20.294516Z",
"published": "2022-03-09T05:26:31Z",

"acquired": "2022-03-05T07:44:38.94697Z",
"published": "2022-03-07T03:34:26Z",

 "acquired": "2022-03-05T07:44:36.47233Z",
"published": "2022-03-06T03:58:26Z",

"acquired": "2022-03-04T07:44:28.091686Z",
"published": "2022-03-05T02:21:31Z",

"acquired": "2022-03-04T07:48:45.835235Z",
"published": "2022-03-05T01:52:01Z",

"acquired": "2022-03-03T08:08:23.881687Z",
"published": "2022-03-06T06:56:31Z",

"acquired": "2022-03-02T07:43:31.594908Z",
"published": "2022-03-02T20:42:22Z",

Investigating a bit further I think the issue is in the filter
{"type": "StringInFilter", "field_name": "quality_category", "config": ["standard"]},

What we recommend you do is:

Best,
Grega

5

Thanks Grega. I did as you suggested and I saw the missing images. However the following images still missing some meta-data (like “visible_percent” and “clear_percent”):

acquired: 2022-03-03T08:08:23.881687Z
published: 2022-03-06T06:56:31Z
…
acquired: 2022-03-05T07:44:36.47233Z
published: 2022-03-06T03:58:26Z
…
acquired: 2022-03-04T07:48:45.835235Z
published: 2022-03-05T01:52:01Z

6

Yes, this is probably the reason, why they are flagged differently with the quality.
As mentioned, we suggest to not rely on these meta-data as they are scene based - each scene is 200km2, so your area of 1 km2 is in most cases unrelated to the meta-data here…

If you use UDM2, you will be able to filtter by AOI.

7

Thanks a lot Grega for your responses.

Is there a way to check for cloud cover inside the AOI only and not in the entire tile?

8

Yes, but you need to change your process a bit:

  • use Subscriptions API to have all observed data available within Sentinel Hub as soon as possible
  • use the meta-data cloud screening option similar to how we describe it for Sentinel-2
9

Thanks a lot.
I will follow your suggestion.

10

Hello!I’m a chinese user.Why can’t I see satellite pictures recently?