DSM Filling

Name: “dsm_filling”

Description

Fill classified values or missing values with one the three avalable methods.

Note

Filling can be configured directly in the input configuration. Each method combines several available filling applications.

Configuration

Name	Description	Type	Available value	Default value	Required
method	Method for hole detection	string	“exogenous_filling”, “bulldozer”, “border_interpolation”		Yes
save_intermediate_data	Save disparity map	boolean		False	No

Method exogenous_filling:

Method “exogenous_filling” fills with altitude of exogenous data (DEM/geoid).

Name	Description	Type	Available value	Default value	Required
classification	Values of classes to fill	List[str]		“nodata”	No
fill_with_geoid	Classes to fill with geoid	List[str]		None	No
interpolation_method	Interpolation method for DEM and geoid resampling	List[str]	“bilinear”, “cubic”	“bilinear”	No

Method bulldozer:

Method “bulldozer” converts the DSM to a DTM and fills the pixels with the output DTM.

Name	Description	Type	Available value	Default value	Required
classification	Values of classes to fill	List[str]		“nodata”	No

Method border_interpolation:

Method “border_interpolation” use the border of every component to compute the altitude to fill.

Name	Description	Type	Available value	Default value	Required
classification	Values of classes to fill	List[str]		“nodata”	No
component_min_size	Minimal size (pixels) of feature to fill	int		5	No
border_size	Size of border used to estimate altitude	int		10	No
percentile	Percentile of border taken for altitude	float		10	No

Note

• If the keyword “nodata” is added to the classification parameter, nodata pixels of the classification will be filled. If no classification is given, nodata pixels of DSM will be filled.

Warning

There is a particular case with the dsm_filling application because it can be called any number of times. Because it is not possible to define three times the dsm_filling in your yaml/json configuration file, you can add an identifier after dsm_filling to differentiate each application :

• dsm_filling.border_interp

• dsm_filling.two

• dsm_filling.with_bulldozer

It is recommended to run bulldozer before border_interpolation in order for border_interpolation to get a DTM. If no DTM is found, border_interpolation will use the DSM. The execution order is determined by the order of the applications in the configuration file.

Example

YAML JSON Copy

---
applications:
  dsm_filling.1:
    method: exogenous_filling
    classification:
    - "8"
    fill_with_geoid:
    - "8"
    save_intermediate_data: true
  dsm_filling.2:
    method: bulldozer
    classification:
    - "6"
    save_intermediate_data: true
  dsm_filling.3:
    method: border_interpolation
    classification:
    - "7"
    save_intermediate_data: true

{
    "applications": {
        "dsm_filling.1": {
            "method": "exogenous_filling",
            "classification": [
                "8"
            ],
            "fill_with_geoid": [
                "8"
            ],
            "save_intermediate_data": true
        },
        "dsm_filling.2": {
            "method": "bulldozer",
            "classification": [
                "6"
            ],
            "save_intermediate_data": true
        },
        "dsm_filling.3": {
            "method": "border_interpolation",
            "classification": [
                "7"
            ],
            "save_intermediate_data": true
        }
    }
}

YAML JSON Copy

---
applications:
  dsm_filling: # run single filling application
    method: exogenous_filling
    classification:
    - "8"
    fill_with_geoid:
    - "8"

{
    "applications": {
        "dsm_filling": {
            "method": "exogenous_filling",
            "classification": [
                "8"
            ],
            "fill_with_geoid": [
                "8"
            ]
        }
    }
}

YAML JSON Copy

---
applications:
  dsm_filling.create_dtm: # the first application uses bulldozer, creating a DTM
    method: bulldozer
    classification:
    - "6"
  dsm_filling.use_dtm: # the second application uses the DTM created earlier
    method: border_interpolation
    classification:
    - "10"

{
    "applications": {
        "dsm_filling.create_dtm": {
            "method": "bulldozer",
            "classification": [
                "6"
            ]
        },
        "dsm_filling.use_dtm": {
            "method": "border_interpolation",
            "classification": [
                "10"
            ]
        }
    }
}