Max depth in output does not correspond with minimum depth parameter

[AnneliesBroeckx]

Hi,

Many thanks to develop this package to detect nested depressions based on a DEM!
I have tested this out with a DEM (resolution 5 m) and height in cm. These were my parameters:

• min_size = 400
• min_depth = 100
• slicing interval = 20

So I have set the minimum depth parameter equal to 100 cm. When I run the script in Python and view the output of the depressions_info.csv file, in order to examine all the detected depressions, I notice that the meximum depth of each polygon does not always correspond with the set minimum depth parameter of 100 cm. Many depressions (376 out of 452) have a maximum depth that is smaller than 100 cm. How is this possible?

I have used the example code (lidar-master\lidar\example.py).

Many thanks for your anwser!

[giswqs]

For 100 cm (1m), it should probably be min_depth = 1

[AnneliesBroeckx]

No, cause the DEM is in centimeters.
I also had done tests with DEM in meter. But then I get similar results (max depth is smaller than set min depth parameter).
Do you know what can be wrong?

[giswqs]

Keep in mind that min_depth will only affect the level-1 depressions. Could you check the out csv and see what levels of those depression with max_depth smaller than min_depth?

[AnneliesBroeckx]

I checked, most of them are level 1-depressions! Do you know what might be cause this error?

[giswqs]

Could you share a small subset of your dataset for debugging?

[AnneliesBroeckx]

I tried to attach here a tif-file, but that filetype is not supported. Then I tried to attach a zip-file, but that didn't seem to work either ("is not included in the list"). How can I send you a DEM tif file?

[giswqs]

Upload to Google Drive and share a link here.

[AnneliesBroeckx]

I have made a subset of my DEM file (see this link). The unit is meters.
Many thanks in advance to check!

[giswqs]

This seems to be a bug. The min_depth parameter has no effect. I will look into it.

[AnneliesBroeckx]

Thank you very much! I look forward to your answer.