DSM merging

Name: “dsm_merging”

Description

Merge multiple DSMs and their associated auxiliary data according to a priority order.

Configuration

Name	Description	Type	Available value	Default value	Required
method	Method for dsm_merging	string	“weighted_fusion”	“weighted_fusion”	No
tile_size	Size of tiling	int		4000	No
filling_priority_list	The filling priority list	list		[6, 1, 4, 5, 3, 2, 7, 0]	No
classif_priority_list	The classif priority list	str		[7, 8, 9, 5, 3, 1, 2, 4, 6, 10, 11, 255]	No

Examples

YAML JSON Copy

merging:
  applications:
    dsm_merging:
      method: "weighted_fusion"
      tile_size: 4000
      filling_priority_list: [1, 6, 4, 3, 5, 2, 7, 0] 
      classif_priority_list: [8, 7, 9, 2, 5, 3, 1, 4, 10, 6, 11, 255] 

{
    "merging": {
        "applications": {
            "dsm_merging": {
                "method": "weighted_fusion",
                "tile_size": 4000,
                "filling_priority_list": [
                    1,
                    6,
                    4,
                    3,
                    5,
                    2,
                    7,
                    0
                ],
                "classif_priority_list": [
                    8,
                    7,
                    9,
                    2,
                    5,
                    3,
                    1,
                    4,
                    10,
                    6,
                    11,
                    255
                ]
            }
        }
    }
}

Priority rules

The DSMs are merged according to a configurable priority order, which can be adjusted in the configuration file. If you request values that are not available in the corresponding filling or classification layers, CARS will fail with an error.

filling_priority_list

value	corresponding filling
6	fill_with_geoid
1	no_edition
4	fill_with_endogenous_dem
5	interpolate_from_borders
3	interpolation
2	fill_with_exogenous_dem
7	other
0	no_data

Following this priority list, only the DSMs and their associated auxiliary data corresponding to the selected filling classes will be merged.

classif_priority_list (using slurp)

value	corresponding classification
7	sea
8	lake
9	river
5	cloud
3	water
1	low_vegetation
2	high_vegetation
4	buildings
6	bare_ground
10	false_positive_buildings
11	background
255	no_data

For the classification, several classification values may be available either for the same filling class or in the case of no_edition. Therefore, The majority classification is selected, and a priority list is used to resolve ties when multiple classifications have the same number of occurrences.