AI Based Smart Agriculture – Leaf Disease Prediction Using Optimized CNN Model


Ashwini A Nadig, M.Tech Scholar,
Dept of ECE, BMSCE, Bengaluru, Karnataka, India
Geetishree Mishra, Assistant professor,
Dept. of ECE, BMSCE, Bengaluru, Karnataka, India.


The Internet of Things is a technology that provides solutions to a number of issues in agriculture. It not only assists in obtaining sensory readings of physical properties but also in connecting those data over the internet utilizing particular protocols. This research focuses on several image processing approaches for leaf disease identification as well as the usage of IOT in farmland for smart farming. The built-in sensors aid in determining the soil’s moisture content, pH level, air temperature, and humidity level. Plant leaf image is Gathered from the field have undergone rigorous preprocessing, which are then subjected to preprocessing based on a Gaussian filter, segmentation, and disease identification using fast R-CNN, faster R-CNN, and Mask R-CNN methods. From the comparative analysis it is identified that mask R-CNN is most suitable method for accurately predicting the plant disease.