Comprehensive AI framework for automated classification, detection, segmentation, and severity estimation of date palm diseases using vision-language models and generative AI

Description

Date palms earn over thirteen billion dollars a year in dry regions, yet pests like the red palm weevil and diseases such as Bayoud and Black Scorch cut yields hard. Checking trees by eye is slow and subjective. Abid Iqbal builds one AI framework that handles the whole diagnostic chain: naming the disease, locating it, outlining the damaged area, and grading its severity.

The pipeline mixes several recent tools. A generative model creates extra training images, CLIP sorts the disease class, and PaliGemma2 finds the spots from text prompts. A pairing of Grounding DINO with SAM 2.1 outlines them without prior examples of the exact class. A vision transformer then scores severity. Across two datasets the system reached 98 percent classification accuracy and detection between 94 and 98 percent. Iqbal presents the unified design as a scalable way to watch palm health.

Catalog record

Source: https://europepmc.org/articles/PMC12753978