Worked examples and output gallery • pointcoral

These examples use files installed with pointcoral, require no internet connection, and write only to temporary directories.

library(pointcoral)

example_dir <- system.file("extdata", package = "pointcoral")
points <- read_cpce_folder(example_dir, image_root = example_dir, recursive = FALSE)
crosswalk <- read_label_crosswalk(file.path(
  example_dir, "pointcoral_example_crosswalk.csv"
))
points_clean <- standardize_labels(points, crosswalk)

Tidy point rows

Original CPCe fields remain beside standardized fields. Pixel coordinates are in the matched image’s pixel coordinate system.

knitr::kable(
  head(points_clean[c(
    "image_id", "point_id", "x_px", "y_px", "raw_label",
    "full_label", "major_category", "class_id"
  )], 10),
  caption = "First ten bundled point annotations."
)

First ten bundled point annotations.
image_id	point_id	x_px	y_px	raw_label	full_label	major_category	class_id
HIW_158_W_U-1	1	19	11	SPO	Sponge	SPONGES (S)	2
HIW_158_W_U-1	2	7	107	S	Sand	SAND, PAVEMENT, RUBBLE (SPR)	8
HIW_158_W_U-1	3	26	165	CALG	Coralline algae	CORALLINE ALGAE (CA)	7
HIW_158_W_U-1	4	69	202	SPO	Sponge	SPONGES (S)	2
HIW_158_W_U-1	5	26	243	SPO	Sponge	SPONGES (S)	2
HIW_158_W_U-1	6	43	305	SPO	Sponge	SPONGES (S)	2
HIW_158_W_U-1	7	2	363	PEYS	Peyssonnelia	PEYSSONNELIACEAE	5
HIW_158_W_U-1	8	74	452	CALG	Coralline algae	CORALLINE ALGAE (CA)	7
HIW_158_W_U-1	9	43	469	SPO	Sponge	SPONGES (S)	2
HIW_158_W_U-1	10	18	541	LOBO	Lobophora variegata	MACROALGAE (MA)	4

Point-count cover

cover <- summarize_points(
  points_clean,
  by = character(),
  class_col = "major_category"
)
knitr::kable(cover, digits = 1, caption = "Pooled bundled-example point counts.")

Pooled bundled-example point counts.
major_category	n	n_points	percent
CORAL (C)	10	200	5.0
CORALLINE ALGAE (CA)	31	200	15.5
MACROALGAE (MA)	18	200	9.0
PEYSSONNELIACEAE	23	200	11.5
SAND, PAVEMENT, RUBBLE (SPR)	74	200	37.0
SPONGES (S)	44	200	22.0

Horizontal bar chart of mean point-count cover across two bundled photoquadrats, led by sand pavement and rubble, then sponges.

The plotted percentages are calculated from 200 demonstration points. Pooling is shown to illustrate output, not to prescribe analysis across sampling units.

Image overlay

Reef photoquadrat with CPCe point circles and short label codes overlaid across the image.

Generate project overlays with write_qc_overlays(). Inspect image orientation, edge points, recognizable benthic features, and the correspondence between point positions and labels.

Point-centered patches

Twelve square reef image patches centered on CPCe points and labeled with example benthic classes.

Patches are square pixel crops. Near an image edge, edge = "skip" omits a crop and edge = "pad" adds black pixels. A patch can contain several benthic types even though its target is inherited from the center point.

Sparse weak-label masks

Sparse mask preview with small colored disks at CPCe point locations and pale gray ignore pixels elsewhere.

The colored disks encode point classes. Pale pixels use ignore_index; they are unknown rather than observed background.

CPCe output workbook raw tabs

read_cpce_output_raw_tabs() handles workbooks with one sheet per image ending in _raw. It keeps within-sheet order as point_index and skips deep_cres_..._raw sheets by default because they use another layout.

raw_tabs <- read_cpce_output_raw_tabs(file.path(
  example_dir, "pointcoral_example_cpce_output_raw_tabs.xlsx"
))
knitr::kable(
  raw_tabs[c(
    "image_name", "point_index", "raw_data",
    "cpce_major_category", "major_category"
  )],
  caption = "Rows from the bundled demonstration workbook."
)

Rows from the bundled demonstration workbook.
image_name	point_index	raw_data	cpce_major_category	major_category
Sample_A	1	SPO	S	SPONGES (S)
Sample_A	2	PEYS	MA	PEYSSONNELIACEAE
Sample_A	3	S	SPR	SAND, PAVEMENT, RUBBLE (SPR)
Sample_A	4	CALG	CA	CORALLINE ALGAE (CA)
Sample_B	1	LOBO	MA	MACROALGAE (MA)
Sample_B	2	SS	C	CORAL (C)
Sample_B	3	P	SPR	SAND, PAVEMENT, RUBBLE (SPR)

For project labels, pass a reviewed crosswalk explicitly. The built-in example mapping should not be assumed to match another CPCe codefile.

Complete output folder

out_dir <- tempfile("pointcoral-example-")
result <- run_pointcoral(
  cpce_dir = example_dir,
  image_root = example_dir,
  out_dir = out_dir,
  crosswalk_path = file.path(example_dir, "pointcoral_example_crosswalk.csv"),
  recursive = FALSE,
  make_patches = FALSE,
  make_masks = FALSE,
  make_qc = FALSE
)

data.frame(
  output = names(result$paths),
  file = basename(unlist(result$paths)),
  row.names = NULL
)
#>                          output                             file
#> 1                    points_raw                   points_raw.csv
#> 2                  points_clean                 points_clean.csv
#> 3             validation_report            validation_report.csv
#> 4               crosswalk_check              crosswalk_check.csv
#> 5                 label_summary                label_summary.csv
#> 6                 image_summary                image_summary.csv
#> 7              transect_summary             transect_summary.csv
#> 8                  site_summary                 site_summary.csv
#> 9     image_summary_clean_label    image_summary_clean_label.csv
#> 10 transect_summary_clean_label transect_summary_clean_label.csv
#> 11     site_summary_clean_label     site_summary_clean_label.csv
#> 12       image_summary_ml_class       image_summary_ml_class.csv
#> 13    transect_summary_ml_class    transect_summary_ml_class.csv
#> 14        site_summary_ml_class        site_summary_ml_class.csv
#> 15                 class_lookup                 class_lookup.csv
#> 16                       labels                       labels.csv
#> 17                 class_lookup                 class_lookup.csv
#> 18                 labels_train                 labels_train.csv
#> 19                   labels_val                   labels_val.csv
#> 20                  labels_test                  labels_test.csv

See Get started for the reasoning behind each step.