docs: to_scl_dilation_mask: document resolution of kernel/erosion?

[soxofaan]

openeo-python-driver/openeo_driver/specs/openeo-processes/experimental/to_scl_dilation_mask.json

Lines 2 to 65 in faeb345

	"id": "to_scl_dilation_mask",
	"summary": "Create a mask for clouds by dilating Sen2Cor sceneclassification",
	"description": "Create a mask for clouds by dilating Sen2Cor sceneclassification. Requires a datacube with only Sentinel-2 L2A 'SCL' band as input. This dilation algorithm removes pixels in the neighbourhood of clouds in a fairly aggressive manner, to avoid any type of contamination. Nevertheless, some outliers can sometimes still remain.",
	"categories": [
	"cubes",
	"ard"
	],
	"experimental": true,
	"parameters": [
	{
	"name": "data",
	"description": "The source data cube containing only the Sentinel-2 SCL band.",
	"schema": {
	"subtype": "datacube",
	"type": "object"
	}
	},
	{
	"name": "erosion_kernel_size",
	"description": "",
	"optional": true,
	"default": 0,
	"schema": {
	"type": "number"
	}
	},
	{
	"name": "mask1_values",
	"description": "",
	"optional": true,
	"schema": {
	"type": "array",
	"items": {
	"type": "number"
	}
	}
	},
	{
	"name": "mask2_values",
	"description": "",
	"optional": true,
	"schema": {
	"type": "array",
	"items": {
	"type": "number"
	}
	}
	},
	{
	"name": "kernel1_size",
	"description": "Size of the first convolution kernel.",
	"optional": true,
	"default": 17,
	"schema": {
	"type": "number"
	}
	},
	{
	"name": "kernel2_size",
	"description": "Size of the second convolution kernel.",
	"optional": true,
	"default": 201,
	"schema": {
	"type": "number"

If I understand correctly this process uses input data at 20m resolution but produces output data at 10m resolution. From the current documentation it is not clear if the provided kernel/erosion sizes are in 10m or 20m pixels

[soxofaan]

Or even better: is it possible to support specifying the sizes in (approx) meter instead of pixels?