Multi-Scale Image Segmentation Model for Fine-Grained Recognition of Zanthoxylum Rust

Zanthoxylum bungeanum Maxim, generally called prickly ash, is widely grown in China. rust the main disease affecting growth and quality of Zanthoxylum. Traditional method for recognizing degree infection mainly rely on manual experience. Due to complex colors shapes areas, accuracy recognition low difficult be quantified. In recent years, application artificial intelligence technology agricultural field has gradually increased. this paper, based DeepLabV2 model, we proposed a image segmentation model FASPP module enhanced features areas. This paper constructed fine-grained dataset. dataset, was segmented labeled according leaves, spore piles, brown lesions. The experimental results showed that effective. rates piles lesions reached 99.66%, 85.16% 82.47% respectively. MPA 91.80%, MIoU 84.99%. At same time, also had good efficiency, which can process 22 images per minute. article provides an intelligent efficiently accurately rust.