README.Rmd

---
title: "coiR"
author: "Edson Silva-Júnior"
date: "`r Sys.Date()`"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

# [coiR](https://github.com/Edbbioeco/coiR)<img src="logo_coiR.png" align = "right" width="150">

> Package to calculate Canopy Openness Index (COI) from canopy images

# Installing package

```{r, eval = FALSE}
require("devtools")

devtools::install_github("Edbbioeco/coiR")
```

# Loading package

`coiR` package is downloaded now, so we can library it. Additionally, we library the following packages:

- [terra](https://rspatial.github.io/terra/): import images files as rasters;

- [purrr](https://purrr-tidyverse-org.translate.goog/?_x_tr_sl=en&_x_tr_tl=pt&_x_tr_hl=pt&_x_tr_pto=tc): create loops for multiple operations;

- [ggplot2](https://ggplot2.tidyverse.org): create elegant graphs for imported images;

- [tidyterra](https://dieghernan.github.io/tidyterra/): visualize RGB raster

```{r}
library(coiR)

library(terra)

library(purrr)

library(ggplot2)

library(tidyterra)
```

# Data

## Importing 

Now, we need to import our data. Images may be shotten photos, as .png, .jpg or .jpeg files. first, we informe images directory (`files`), and import them using `terra::rast()` function for every image, throught a loop with `purrr::map()` function. Our images (`images`) are setted as a list class object.

```{r}
files <- paste0("cropped-images/imagem", 1:4, ".png")

files

images <- purrr::map(files,
                     terra::rast)

names(images) <- paste0("cropped-images/imagem", 1:4, ".png")

images
```

## Visualizing

Next, lets visualize every image, using a `purrr::map()` loop, through `ggplot` and `tidyterra::geom_spatraster_rgb()` function.

```{r}
purrr::map(images, function(data){ggplot() + tidyterra::geom_spatraster_rgb(data = data) + theme_void()})
```

# Canopy Openness Index 

## Isolating a single image

For our exemples, lets work in two ways: run for a single and run for multiple images. Lets set first image as a single image.

```{r}
single_image <- images[[1]]
```

## Crop image

Usely, canopy images are shotten photos, square images. For our analysis, we need to crop images into circles. We use `coiR::coir_crop()` |function.

```{r}
single_image |> 
  coiR::coir_crop()
```

And we also can analyse multiple images from an one shot, using `purrr::map()` loop.

```{r}
purrr::map(images, coiR::coir_crop)
```

## Binarize images

Our next step is to binarize our images. We use `coiR::coir_binarize()` function. To avoid replot canopy image, we use `plot = FALSE` argument at `coiR::coir_crop()` function. To facilite our analysis, lets use pipe (`|>`) to conect functions output.

```{r}
single_image |> 
  coiR::coir_crop(plot = FALSE) |> 
  coiR::coir_binarize()
```

As previously made, we can binarize multiple images, making a function in our `purrr::map()` loop.

```{r}
purrr::map(images, function(images){coiR::coir_crop(data = images,
                                                    plot = FALSE) |>
    coiR::coir_binarize()})
```

## Index

Finally, we get COI index. We use `coiR::coir_index()` for get that value. To avoid replot cropped and binarized images, we set `plot = FALSE` for both `coiR::coir_crop()` and `coiR::coir_binarize()` functions.

```{r}
single_image |> 
  coiR::coir_crop(plot = FALSE) |> 
  coiR::coir_binarize(plot = FALSE) |> 
  coiR::coir_index()
```

As previously made, we can also do for multiple images, making a function in `purrr::map()` function.

```{r}
purrr::map(images, function(images){coiR::coir_crop(data = images,
                                                    plot = FALSE) |>
    coiR::coir_binarize(plot = FALSE) |> 
    coiR::coir_index()})
```

Additionally, we can set all COI values into a vector, to make a dataframe. First, we make a null vector (`vector_index <- c()`). Next, we build a function, hyper declaring a new `vector_index` with `<<-` instead `<-`. For more values details, we set `round = 3` argument in `coiR::coir_index()` functions, to set decimal places. All in a `purrr::map()` loop.

```{r}
vector_index <- c()

multiple_index <- function(images, verbose = FALSE){
  
  index <- images |> 
    coiR::coir_crop(plot = FALSE) |>
    coiR::coir_binarize(plot = FALSE) |> 
    coiR::coir_index(round = 3)
  
  vector_index <<- c(vector_index, index)
  
}

purrr::map(images, multiple_index)

vector_index
```

Finally, we make a dtaframe with those values.

```{r}
df_index <- tibble::tibble(id = images |> names(),
                           Index = vector_index)

df_index
```