Image dataset of infected date palm leaves by dubas insects

Published: 26 April 2023| Version 2 | DOI: 10.17632/2nh364p2bc.2
Contributors:
Abdullah Mazin, Haider Almayaly

Description

The palm leaf images were categorized based on their health status and the presence of insects, resulting in four categories: healthy, infected with bugs only, infected with honeydew only, and infected by mixed insects and honeydew. Images of leaves infected with insects depict a range of insect life cycle stages, from the third generation of nymphs to the adult stage in the fifth nymph stage. Two drone cameras were employed to capture the images, resulting in a dataset of 3000 images, with 800 per non-bug category and 600 for the bug category. The dataset is valuable for assessing infestation severity, estimating insect populations, and determining the extent of damage.

Files

Steps to reproduce

To obtain images of infected palm leaves, a survey of agricultural lands infested with Dubas insects was conducted with the assistance of an agricultural guide. The survey covered a period of the insect life cycle, including spring and autumn generations. Infected palms were identified during land scanning, and the drone captured images of the leaves from a distance of 1-2 meters. In autumn, small bugs with eggs were sometimes present on the leaves, while clear bugs appeared in spring. Honeydew was observed on the leaves at the end of spring and summer. The resulting images were categorized into four folders based on the season of capture and were screened for noise, shadow, and dust. All imaging was performed with the consent and knowledge of the orchard owners. The dataset provides valuable information for monitoring and managing infestations in agricultural lands.

Institutions

University of Karbala

Categories

Computer Science, Artificial Intelligence, Insect, Machine Learning, Artifact Detection, Deep Learning

Licence