insects

Dataset Overview

Dataset Name

• Name: insects

Dataset Type

• Type: Object Detection Dataset

Dataset Description

The dataset aims to improve the YOLOv11 crop pest detection system, specifically for training and evaluating model performance on crop pest identification. It encompasses ten pest categories: armyworm, legume blister beetle, red spider, rice gall midge, rice leaf roller, rice leafhopper, rice water weevil, wheat aphid, white‑backed plant hopper, and yellow rice borer. Accurate identification and timely control are critical due to the severe threat these pests pose to agricultural production.

Dataset Composition

• Number of Classes: 10
• Class Names:
  • army worm
  • legume blister beetle
  • red spider
  • rice gall midge
  • rice leaf roller
  • rice leafhopper
  • rice water weevil
  • wheat phloeothrips
  • white backed plant hopper
  • yellow rice borer

Dataset Scale

• Number of Images: 995

Dataset Characteristics

• Constructed from extensive field collection and annotation, ensuring authenticity and diversity.
• Each pest class is precisely annotated by specialists, covering various growth stages and environments, enhancing model generalization.
• Diverse samples enable the model to adapt to different lighting conditions, background clutter, and pest postures, improving detection accuracy and robustness.

Dataset Uses

• Training and evaluating an improved YOLOv11 model for crop pest detection tasks.
• Leveraging high‑quality annotations with YOLOv11’s advanced features for iterative training to optimize model parameters and architecture.
• Images are split into training, validation, and test sets for comprehensive performance assessment.

Dataset Download

Project Dataset Download Link