Detection and infected area segmentation of apple fire blight using image processing and deep transfer learning for site-specific management

Md. Sultan Mahmud, Long He, Azlan Zahid, Paul H. Heinemann, Daeun Choi, Grzegorz Krawczyk, Heping Zhu

Abstract: Highlights•A Mask R-CNN based fire blight detection and segmentation system was introduced.•ResNet-50 and ResNet-101 backbone were used for feature extraction.•The transfer learning approach was used to reduce data requirement.•Vegetation indices were calculated to analyze the individual index sensitivity to apple fire blight.•Results suggest that the system can detect fire blight in complex orchard conditions.

External IDs:dblp:journals/cea/MahmudHZHCKZ23