Published December 7, 2023 | Version 1.0
Dataset Open

Maize Phosphorus Leaf Deficiency (MPLD) Database | Compact Scientific Camera (original-processed)

Authors/Creators

  • 1. ROR icon Universidad EAFIT

Description

This database presents samples of maize leaves placed on a withe background, representing three levels of phosphorus deficiency: complete absence of the nutrient (labeled -P), half dose of the required phosphorus for normal plant development (-P50), and complete supply (C).

Its composed of two folders:

  • Original_dataset: 722 jpg images of 1280 x 1020 pixels size divided into '_C', '-P' and '-P50' folders for class labels.
  • Processed_dataset: 2433 png images of 224 x 224 pixels size divided into '-C', '-P' and '-P50' folders for class labels.

Methods

Data collection

Image acquisition was done by searching a diversity of conditions, although the plants were grown in laboratory conditions. In this way, photographs included the growth stages of seedling, jointing, and flowering. The images were acquired inside a plastic shed, involving natural illumination, and five consecutive images were taken by plant. 

Acquisition device was a compact scientific camera (ThorLabs DCC1645C-HQ).

Image preprocessing and augmentation

The original images underwent a preprocessing stage in terms of cropping and resizing by two methods done simultaneously: 

  1. Cut to central square: Image cropped to a central square.
  2. Cuadrant division: Square image divided into four equal parts.

All cropped images were resized to 224x224 pixels size.

As cuadrant division method can generate sub-images with black content or a minimum portion of the leaf, and these images can confuse the algorithm, a suffix is added to the images cosidered as empty on the Processed_dataset. 

Notes

Composition of Processed_dataset folder

Each image has a filename as 'S<xx>-T-<class>-<yy><_n>

Were

  • <xx> goes from 05 to 10 and represents the week of maize growth at the time image was collected.
  • <class> could be C, P or P50
  • <yy> is a unique id for each image and
  • <n> Could be present or not, indicating the number of the cuadrant obtained from the Cuadrant Division process, and goes from 0 to 3
  • In addition, some images have the suffix '_EMPTY' which leads to a blank image or one that does not have enough leaf content.

For each image taken from a single leaf, there are 5 processed images in this folder.

Files

Files (208.8 MB)

Name Size Download all
md5:9228abbf88abee25ed55bf91581a2ac3
43.7 MB Download
md5:9de0ba2a2da1d14f91b69b7b8e10bfd0
165.1 MB Download

Additional details

Dates

Collected
2022-05-13