Abstract:
This study proposes an autonomous system for identifying and classifying mango trees using the DenseNet-201, MobileNet, ResNet50, ResNet-50 Transfer Learning model. Mango tree species must be accurately identified and classified in order to maximize agricultural methods, enhance resource management, and guarantee sustained mango production. Due to human subjectivity, traditional tree identification techniques that depend on the manual observation of leaf attributes are time-consuming, labor-intensive, and frequently inconsistent. This paper presents an automated system that can recognize and categorize different kinds of mango trees using photos of their leaves in order to overcome these difficulties. To improve model performance, a specially prepared dataset with 1,200 photos of mango leaves from seven different species was made. The ResNet-50 model demonstrated its efficacy in classifying mango tree species with an impressive accuracy score of 96.85% after being modified for this task. The study's findings demonstrate how computer vision techniques can be used to automate the categorization process, eliminating the need for human labor and facilitating accurate, large-scale mango tree identification.
