Detecting Leaf Diseases Using Centralized CNN'S with Efficient R-CNN and Mask R-CNN

Plant diseases have emerged as a significant concern, as they can drastically reduce both the quality and quantity of agricultural products. The research aims to identify plant leaf infections early on, enabling timely treatment to prevent the spread of diseases to other plants. The objective is to determine the severity of leaf diseases, allowing for targeted treatment based on the level of severity. Additionally, the severity of leaf diseases can help predict crop losses. This work is to improve the accuracy of leaf detection and measurement of morphological traits by Region Convolutional Neural Network (RCNN), Network namely state of the Mask region art. Experimental results verify that our automated approach effectively segments leaves from different plant families. Measured leaves of comparable size have an average relative error of about 4.6 percent for leaf length and approximately 5.7 percent for leaf width when compared to manual measurements by expert ecologists and botanists.