PhenoSpectra: Multispectral Data Analysis and Visualization

Description

Provides tools for processing, analyzing, and visualizing spectral data collected from 3D laser-based scanning systems. Supports applications in agriculture, forestry, environmental monitoring, industrial quality control, and biomedical research. Enables evaluation of plant growth, productivity, resource efficiency, disease management, and pest monitoring. Includes statistical methods for extracting insights from multispectral and hyperspectral data and generating publication-ready visualizations. See Zieschank & Junker (2023) doi:10.3389/fpls.2023.1141554 and Saric et al. (2022) doi:10.1016/J.TPLANTS.2021.12.003 for related work.

Author(s)

Maintainer: Medhat Mahmoud medhat.mahmoud@bayer.com

Other contributors:

• Arianna Del Papa arianna.delpapa@bayer.com [contributor]

• Melina Rampelmann melina.rampelmann@bayer.com [contributor]

• Patrick Dohle patrick.dohle@bayer.com [contributor]

.onLoad Function

Description

Automatically runs when the package is loaded and sets up the environment.

Usage

.onLoad(libname, pkgname)

Feature Selection for Spectral Data

Description

This function filters healthy vs diseased samples, selects the most discriminative spectral variables, applies FDR correction, and exports the results.

Usage

feature_selection(
  file_path,
  output_path = "selected_features.xlsx",
  fdr_threshold = 0.01
)

Arguments

Value

A data.table containing selected spectral variables.

Examples

# Create mock spectral data
library(openxlsx)
mock_data <- data.frame(
  treatment = sample(0:1, 100, replace = TRUE),
  var1 = rnorm(100),
  var2 = rnorm(100),
  var3 = rnorm(100)
)
temp_file <- tempfile(fileext = ".xlsx")
write.xlsx(mock_data, temp_file)

# Perform feature selection
output_path <- tempfile(fileext = ".xlsx")
selected_features <- feature_selection(temp_file, output_path, fdr_threshold = 0.01)
head(selected_features)

Predict Spectral Disease Severity (SDS)

Description

This function predicts Spectral Disease Severity (SDS) using a standard linear regression model (lm()). It automatically handles column names with special characters by using backticks and constrains predictions to the range ⁠[0, 100]⁠.

Usage

predict_SDS(cleaned_data, sf_test, fixed_effects = NULL)

Arguments

Value

A dataframe with predicted SDS values for all treatments, constrained between 0 and 100.

Examples

# Create mock spectral data
library(openxlsx)
cleaned_data <- data.frame(
  treatment = sample(0:1, 100, replace = TRUE),
  var1 = rnorm(100),
  var2 = rnorm(100),
  var3 = rnorm(100),
  Scan.date = sample(
    seq.Date(
      from = as.Date('2023-01-01'),
      to = as.Date('2023-12-31'),
      by = 'day'
    ),
    100
  ),
  Scan.time = format(Sys.time(), "%H:%M:%S")
)

Perform QA/QC on spectral data while preserving original column names

Description

Perform QA/QC on spectral data while preserving original column names

Usage

qaqcs(
  file_path,
  output_path,
  handle_missing = "NA",
  handle_outliers = "NA",
  group_by_col = "treatment"
)

Arguments

Value

A list with cleaned data and a summary table

Examples

library(openxlsx)
# Create mock raw data
raw_data <- data.frame(
  treatment = sample(0:1, 100, replace = TRUE),
  var1 = rnorm(100),
  var2 = rnorm(100),
  var3 = rnorm(100)
)

# Save mock data to a temporary file
raw_data_file <- tempfile(fileext = ".xlsx")
output_file <- tempfile(fileext = ".xlsx")

write.xlsx(raw_data, raw_data_file)

# Run QA/QC with missing values imputed and outliers removed
cleaned_result <- qaqcs(
  file_path = raw_data_file,
  output_path = output_file,
  handle_missing = "impute",
  handle_outliers = "remove",
  group_by_col = "treatment"
)
head(cleaned_result$cleaned_data)

Read and merge spectral data using data.table exclusively

Description

Read and merge spectral data using data.table exclusively

Usage

reads(directory, pattern, output_path)

Arguments

Value

A merged data.table

Examples

library(data.table)
# Create mock data files with all required columns
mock_data1 <- data.frame(
  treatment = sample(0:1, 50, replace = TRUE),
  var1 = rnorm(50),
  var2 = rnorm(50),
  Scan.date = sample(
    seq.Date(
      from = as.Date('2023-01-01'),
      to = as.Date('2023-12-31'),
      by = 'day'
    ),
    50,
    replace = TRUE
  ),
  Scan.time = format(Sys.time(), "%H:%M:%S"),
  timestamp = Sys.time()  # Add timestamp column
)

mock_data2 <- data.frame(
  treatment = sample(0:1, 50, replace = TRUE),
  var1 = rnorm(50),
  var2 = rnorm(50),
  Scan.date = sample(
    seq.Date(
      from = as.Date('2023-01-01'),
      to = as.Date('2023-12-31'),
      by = 'day'
    ),
    50,
    replace = TRUE
  ),
  Scan.time = format(Sys.time(), "%H:%M:%S"),
  timestamp = Sys.time()  # Add timestamp column
)

# Save mock data to temporary CSV files
temp_dir <- tempdir()
file1 <- file.path(temp_dir, "input_file1.csv")
file2 <- file.path(temp_dir, "input_file2.csv")
fwrite(mock_data1, file1)
fwrite(mock_data2, file2)

# Run the reads() function on mock CSV data
merged_data <- reads(
  directory = temp_dir,
  pattern = "input",
  output_path = tempfile(fileext = ".csv")
)
head(merged_data)

Setup Environment for PhenoSpectra

Description

Automatically installs and loads all required packages.

This function installs missing packages required by the PhenoSpectra package.

Usage

setup_environment()

setup_environment()

Value

NULL. This function installs required packages if missing and loads them.

Examples

# Example usage:
setup_environment()  # Automatically installs and loads required packages
# Run to manually install missing packages
setup_environment()

PhenoSpectra Internal Setup

Description

Handles automatic environment setup and global variable registration when the package is loaded.

Examples

# Example usage:
# Automatically run when the package is loaded, but can be manually called if needed
setup_environment()