A novel model for precise segmentation and disease diagnosis of economic forest images in complicated backgrounds

Liang Qi

doi:10.1016/j.compag.2026.112279

A novel model for precise segmentation and disease diagnosis of economic forest images in complicated backgrounds

2026-08-06

0

OA

AI

L

Liang Qi

C

Chi Zhou

B

Bin Wu

Q

Qiaolin Ye

M

Mingguang Li

W

Weijun Xie

S

Shuaishuai Zhao

K

Kai Zhang

DOI：10.1016/j.compag.2026.112279

Original

Original request for help

Save

Abstract

En 中文

• A high-resolution UAV ginkgo canopy dataset is constructed for disease classification. • A novel Haar-UNet is proposed to enhance segmentation in dense and complex backgrounds. • A lightweight Haar-DiseaseNet is developed for multi-type ginkgo leaf disease identification. • The two-stage framework adapts well to dense planting and complex field backgrounds.

Keywords:

Haar-UNet

Semantic segmentation

Disease identification

Drone photography

Ginkgo biloba

A novel model for precise segmentation and disease diagnosis of economic forest images in complicated backgrounds

Abstract

Journal

Organization

Cited Papers

Citing Papers