I am using the Ocean Object Detection algorithm in both SNAP11 (I have used it
many times successfully) and in SNAP12 with the same data set. In SNAP11 I was successful, but in SNAP12 the adaptive threshold algorithm (as graph builder and within the Ocean Object Detection workflow) finishes very quickly and is writing an image, but no detections (masks, vectors or CVS file) are created. Could you please take a look at it?
Could you provide more information (e.g. product names, processing steps or graph and parameters used) regarding your processing so that we can reproduce the problem? Thank you.
Sure, I used the example from RUS Copernicus OCEA01, Gulf of Trieste Italy with the given S1 GRDH image of 2016-10-09. I created a subset, applied the orbit file and then the SNAP 12 Ocean Object Detection workflow as provided in the graph builder (SRTM3 land masking, radiometric calibration to sigma-0, adaptive thresholding, object discrimination using the default settings in according with the above mentioned tutorial.
The output has the correct name but does not include any detection. Also the process finishes way too quickly (47 sec).
If I do the steps manually and with the same settings, I had to do the adaptive Thresholding in a small model: read subsetted, masked, calibrated data, apply thresholds as given in the tutorial, write to file. in Vectors and in Bands, no detections are included. The output only is a rewrite of the sigma-0 data of both bands. No error message is issued…
Thanks for your help.
A Jira ticket (Jira) has been created to track the problem. We will look into it. Thank you for pointing out the issue.
The problem has been fixed and the fix should be in the next release. Thank you.
1 Like