GRD assets available in Google Earth Engine

Microwave Toolbox
#62

Can you provide the asset ids or the area that you are looking at?

The general answer is that sometimes there are duplicate assets, and EE is not always doing a great job deduplicating them. Fortunately, scihub just started providing their full asset catalog in CSV files, so I’m going to use that for proper asset cleanup.

#63

Actually I looked at the dataset I found and…yes! There are two images for some days (see image attached).

image.png758×101 21.1 KB

I wonder why there are two overpasses… I thought it was one during the ascending orbit and one during the descending orbit pass…

#64

What do you mean with “provide the assets”? Anyway, I’m looking at an area in Thailand.

Although I have another question about the time series chart generation in GEE: in the script for the chart generation I set up the option for the mean generation… but actually it seems that it doesn’t work!

// Create an image time series chart.

var chart = ui.Chart.image.seriesByRegion({
imageCollection: sentinel1,
regions: merge,
reducer: ee.Reducer.mean(),
scale: 30,
xProperty: ‘system:time_start’,
seriesProperty: ‘CANE_TYPE2’
});

#65

ok I got it, I mean why I have two overpasses…

image.jpg856×786 245 KB

#66

These are 2 adjacent frames from the same orbit (“overpass”), because your area straddles both. Logically, they are from the same date, 25 seconds apart, which is the typical time slice for a GRD frame. There is no duplication.

1 Like
#67

@simonf, are you still only updating the S1_GRD once a week? I’m trying to find a solution somewhere with mosaic’ed Sentinel-imagery that is updated as soon as images appear over at ESA.

#68

We are fairly close to making the updates daily.

#69

Nice! I just saw that the latest image was from the 17th. I’m using it to detect sea ice and therefore it’s crucial for me that the images are as fresh as possible.

#71

Capture.PNG722×485 604 KB

How do I correct this error (red circle) in GEE? I want to perform flood mapping but when I mosaic the scenes, the black edges of the images also appear.

#72

There are known issues with border noise in recently processed images: see GRD-Border-Noise and Thermal noise removal are not working anymore since March 13. 2018 and No data value after standard GRD processing

Once these are resolved, we’ll reprocess the affected S1 scenes

#73

How’s your work getting the updates daily? Right now I can see two day old S1 on GEE.

#74

The updates of the top-level computed dataset are now daily rather then weekly. But yes, it still takes about two days for a scene to appear in EE. We are currently working on making Sentinel-2 updates be faster then daily (S2 processing is simpler than S1). Once that is stable, we’ll start improving S1 latency.

#75

Dear all,
I’m new both in SAR and in GEE, apparently, I’m bad also in scripting. But I want to do Sentinel 1 time-series analysis for forest biomass. SNAP didn’t support doing batch for many scene’s preprocessing so I had to switch to GEE. Here I have some problems and I would appreciate any help.

  1. the amount of scenes within the given parameters for the whole year are 60, but the elements are 72. After stacking the amount of bands are finally 72 and they contain also those 12 elements without value (called ‘masked’). How I can get rid of those elements? (i’m attaching a screenshot)
  2. while using Refined Lee filter, is it mandatory to apply it for the each single scene or can I use it after one month averaging (which I’ll be using later for time series analysis)?
  3. Metadata says in DEE the terrain radiometric calibration for S1 is done with sigma naught output. Is there a way to get the gamma naught or should I start preprocessing the very raw data on my own?
    Here I’m posting my code:
    https://code.earthengine.google.com/e0a867c01a79ed6f8355562b28edd26d
    Screenshot_5
#76

  1. Please make users/armkhudinyan/study_area publicly readable. If I use ‘geometry’ instead, there are always 72 bands: https://code.earthengine.google.com/9f263f5f6271c14e1714881d24e79c7f

  2. I suspect Refined Lee Filter would behave differently on a composite because speckling would not have the same characteristics, but I’m not a specialist. If you don’t get a response here, post in the EE user group: https://groups.google.com/forum/#!forum/google-earth-engine-developers

  3. This is a user-posted example of converting to gamma naught:
    https://code.earthengine.google.com/5b1b6d689e5098473708004ec227a714 (from https://groups.google.com/d/topic/google-earth-engine-developers/3-q0TEwa-Tk/discussion)

#77

Thanks a lot @simonf for the nice response.

I shared the study area public. The thing is shows 72 elements in Console, but when you go on Inspector and click on the study area map it shows 60 images and 72 elements, and as it’s on the screenshot, some of the elements are just without value and called “masked”.

#78

Inspector just skips images that have no data for the point you click on (if you click on areas at the bottom, there are only 5 images in some places).
Generally you don’t need to get rid of masked-out pixels, as they are handled correctly (eg, they won’t skew per-pixel averaging). Are they creating some specific problems?

#79

I’m still not sure if it can make problems later on, but it’s good to know that they don’e affect on the averaging. Just was weird that those 60 images are constantly available within my study area (’‘table’’ one).
Thanks for the help and the tips.

#80

ooooooo yeah !!
thanks

#81

thanks for sharing

#82

hi team,
where we can get the sample google earth engine codes for the identification of land displacement & subsidence using sentinel 1.

thank you in advance.