CNES/cars

CARS runs without any error but does not write dsm.tif file on disk

Closed this issue · 5 comments

zoepapirer commented

Hello,

I'm new to using CARS. I've tried computing a DSM with the sample data of Gizeh that you provided and it worked fine, I managed to compute both the dsm.tif and clr.tif files, thanks a lot for this example.

I'm now trying to compute a DSM using stereo Pléiades data. Based on the documention available and previous issues, I've organised the configuration file as follow :

{
    "inputs": {
        "sensors" : {
            "one": {
                "image": "/home/data/6798639101/IMG_PHR1B_P_001/IMG_PHR1B_P_201804150756576_SEN_6798639101-1_R1C1.TIF",
				"geomodel": "/home/data/6798639101/IMG_PHR1B_P_001/DIM_PHR1B_P_201804150756576_SEN_6798639101-1.XML"
            },
            "two": {
                "image": "/home/data/6798640101/IMG_PHR1B_P_001/IMG_PHR1B_P_201804150756295_SEN_6798640101-1_R1C1.TIF",
				"geomodel": "/home/data/6798640101/IMG_PHR1B_P_001/DIM_PHR1B_P_201804150756295_SEN_6798640101-1.XML"
			}
        },
        "pairing": [["one", "two"]],
        "initial_elevation": "/home/data/fabdem_large.tif"
    },
    "output": {
        "out_dir": "/home/data/outresults_1"
    },
    "orchestrator": {
        "mode": "local_dask",
        "nb_workers": 4,
		"max_ram_per_worker": 12000,
		"activate_dashboard": true,
		"walltime": "12:00:00",
		"profiling": {
			"mode": "memray",
			"activated": true
		}
    }
}

The process runs smoothly, without encountering any errors (only pausing and resuming at times, based on memory usage criterias). However, after completing the "Processing Tiles: [ color , dsm ]" step, the process stops without writing the dsm.tif and the clr.tif file in the output folder. I've included both the content.json and used_conf.json generated during the process.

Since there aren't any error message after the process, I'm confused about this issue. Have you ever encountered it ? Why would the writing fail ?

Thank you so much in advance for your help.

content.json
used_conf.json

dyoussef commented

Hi @zoepapirer,

We are currently experiencing some difficulties with the disparity intervals to be explored. We're releasing version 0.7.4 with local disparity intervals which may help solve the problem.

I'll try to reproduce the problem on my side and suggest a solution to successfully generate your DSM.

I'll keep you posted,
David

zoepapirer commented

Hi @dyoussef,

Thank you very much for your answer and your consideration !

I'll wait for your feedback.
Zoé

zoepapirer commented

Hi, and Happy New Year !

I just wanted to give you an update on this issue.

I've tried running CARS on the same dataset but with the latest version of CARS.
Again, the process runs smoothly, ending with the "CARS has successfully completed the pipeline" message, however, there are still no dsm.tif or color.tif files generated.

The files dem_mean.tif, dem_max.tif and dem_min.tif are generated.
I've also tried running the latest version of CARS on the Gizeh data, and got both the dsm.tif and clr.tif files.

You can find the used_conf.json file attached to this message, the parameters are the same as the ones from the test I did back in November. Am I doing something wrong ?

Thank you again for your help.
Zoé

used_conf.json

dyoussef commented

Hello @zoepapirer,

A new version of CARS has been released: https://pypi.org/project/cars/0.7.5/
I managed to generate the DSM with this new version.

Here the DSM:
sertit_dsm

Here the CLR:
sertit_clr

Here the my config file:

{
    "inputs": {
        "sensors": {
            "one": {
                "image": "PWH-7903_2023-12-18_08-54-19_577/IMG_PHR1B_P_001/DIM_PHR1B_P_201804150756295_SEN_PWH-7903_2023-12-18_08-54-19_577-001.XML",
                "geomodel": "PWH-7903_2023-12-18_08-54-19_577/IMG_PHR1B_P_001/RPC_PHR1B_P_201804150756295_SEN_PWH-7903_2023-12-18_08-54-19_577-001.XML"
            },
            "two": {
                "image": "PWH-7903_2023-12-18_08-54-19_578/IMG_PHR1B_P_001/DIM_PHR1B_P_201804150756576_SEN_PWH-7903_2023-12-18_08-54-19_578-001.XML",
                "geomodel": "PWH-7903_2023-12-18_08-54-19_578/IMG_PHR1B_P_001/RPC_PHR1B_P_201804150756576_SEN_PWH-7903_2023-12-18_08-54-19_578-001.XML"
            }
        },
        "pairing": [["one", "two"]]
    },
    "output": {
        "out_dir": "out"
    },
    "orchestrator": {
        "mode": "multiprocessing",
        "nb_workers": 100
    }
}

Regards,
David

zoepapirer commented

Hello @dyoussef,

I'm sorry for the very late reply but I have some great news : I can confirm that it worked fine on my computer too.

image

Thank you for your support !
This issue can now be closed.

Regards,
Zoé

  • 🎉1