Enhancing crop disease classification and severity assessment through Fuzzy Rank-based ensemble learning approach

Yasmine Medjadba, Hamza Drid, Xianchuan Yu

Abstract: Plant diseases affect crop yields, making accurate detection and severity assessment vital. This study proposes a framework combining pre-trained CNNs with a fuzzy rank-based ensemble for improved classification. Three CNNs are integrated, with predictions guided by confidence assessments. For diseased images, the model localizes affected areas and uses the Analytic Hierarchy Process (AHP) to assess severity as mild, moderate, or severe. Results show the method outperforms existing approaches in both detection and severity estimation.

External IDs:dblp:journals/sivp/MedjadbaDY25