Generate-crown-labels • canObsR

Summary

The ‘Generate crown labels’ article describes the process for (1) extracting crown images from the othomosaics and (2) add labels to each images (Fig 1).

Fig 1 : Generate crown labels process

Requirements

This step-by-step process requires :
- crown delineation file compatible with the canObsR functions. Check the Create compatible crowndelineation file article for more details.
- Orthomosaics time-serie. Check the Generate and align orthomosaics article for more details.

To run the following code you can download and unzip ‘canObsR_data2.zip’.

Step-by-step exemple

Extract crown images

Set the mosa_path to the time series orthomosaics folder, the crownFile_path to the crown delineation file, the site name, the Tmpfiles_directory if you want to store the temporary files in a specific folder and N_cores the number of cores you want to allocate for the parallelized functions*.

Note* : If you are using you own data, the function extract_crownsImages() can produce a lot of temporary files and need several GB of memory depending on the size, the resolution and the number of mosaics. You can specify where you want to store the temporary files.

############## Load libraries --------------------------------------------------

library(canObsR)
library(tidyverse)
library(sf)
library(openxlsx)

############## Set parameters --------------------------------------------------

canObsR_data2_folder = 'XXXXX/canObsR_data2'

# Path to the time-serie main directory
mosa_path = file.path(canObsR_data2_folder, '3_orthomosaics_aligned')

# Path to the crown delineation file
crownFile_path = file.path(canObsR_data2_folder, 'crowns.gpkg')

# Path to the crown delineation file
imgs_folder = file.path(canObsR_data2_folder, '4_crowns_img')

# Path to store the temp files
Tmpfiles_directory = NULL # if NULL it use the default folder for tmp file. You can specify any folder

# Dates of time-serie images
dates <- c("20220427", "20220511", "20220525")

# Number of cores used for the paralelized functions
N_cores = 10

# Site name
site = 'Bouamir'

# Path to the mosaics
mosa_path <- list.files(mosa_path, full.names = T)

Check the crown file compatibility (Fig 2), (more details here)

crownFile <- read_sf(crownFile_path)

check_crownFile(crownFile)

Fig 2 : Crown file compatibility

Extract individual crown images to the ‘canObsR_data2/4_crowns_img’ folder. For the test dataset it can take fiew minutes.


extract_crownImages(
   path_images = mosa_path,                                         # Path to mosaics
   path_crowns = crownFile_path,                          # Path to  crowns file
   out_dir_path =  imgs_folder,                         # Path to output folder
   tempdir_custom = Tmpfiles_directory,             # Path to tmp file directory
   dates = dates,                                  # dates of time series images
   N_cores = N_cores,                                          # Number of cores 
   width = 720,                                           # By default it is 720 
   height = 825                                           # By default it is 720 
)

Create labeling file

create_labelingFile(
   crownFile = crownFile,          # Path to  crown file
   site = site,                           # Site name
   dates = dates,                         # Dates of the RGB survey
   out_dir_path = canObsR_data2_folder)   # Path to output folder

Do labeling with the app


shiny_labels(
   labelingFile = file.path(
      canObsR_data2_folder, 
      paste0(site, "_WideLabels_", format(Sys.Date(), "%Y%m%d"), ".xlsx")
      ),
   imgFolder = imgs_folder
)