Deep Transfer-Learning Assisted Healthy/Disease Strawberry Leaf Images Classification

A lot of different fields, including agriculture, use automatic data handling. Using computer algorithms to handle agricultural data has become an important way to keep an eye on crops at different times. The goal of this study is to create a tool that can use deep transfer learning (DTL) to find plant diseases by looking at pictures of their leaves. This plan has three steps: collecting and preprocessing images, reducing features with DTL, and classifying images into two groups using five-fold cross validation. A small group of well-known, lightweight models that support transfer learning are used to test how well the suggested tool works. The proposed classification job is carried out using the SoftMax classifier, and the accuracy of the detections shows that this method works. As proof in the real world, pictures of strawberry leaves were used, and the method helped achieve detection accuracy of 100% when ResNet18 was used and over 97% when MobileNet was used. These results show that the suggested tool with the DTL does a good job on the picked database.