Cactus dataset
Description
The global climate is currently experiencing rapid changes that have significant impacts on various regions. Consequently, there is a growing inclination towards cultivating non-traditional crops, such as the prickly pear. People are now exploring the diverse health and aesthetic benefits of the prickly pear, beyond its conventional applications in livestock feed and farmland protection. Given this mounting interest, it is imperative to develop enhanced methods for the early detection and treatment of cactus diseases. In light of the limited availability of cactus databases, we have undertaken the task of compiling a comprehensive collection of images portraying both healthy cacti and those infected with the cochineal insect. The intended purpose of this collection is to facilitate advancements in the fields of deep learning and computer vision.
Files
Steps to reproduce
The data was collected in two phases, utilizing a previously unpublished private database.All photos were taken in Sidi Bouzid, Tunisia. We conducted a comprehensive investigation and divided the data collection process into two distinct phases. During the first phase, we visited the cactus fields in the Labid area that had been devastated by cochineal insects. Equipped with a high-quality Canon EOS 800D camera, the team captured a total of 550 photographs on February 3, 2024. Each image had a resolution of 2400 pixels by 1600 pixels, with a pixel density of 72 pixels per inch. Moving on to the next phase, our attention turned to the intact cactus fields in the Al Galel area. Over the course of two days, on February 11 and 12, 2024, we used a smartphone to take 550 photographs of healthy cacti. These images displayed a higher resolution of 3000 pixels by 4000 pixels, while maintaining a constant pixel density of 72 pixels per inch. The images were first sorted, with duplicates and non-expressive images being deleted. Afterward, the images were standardized and saved in PNG format. As a result, the final database consisted of a total of 628 images, with 343 representing injured cacti and 285 representing healthy cacti.