Image Based Plant Disease Detection Using ML Algorithm

Abstract

The suggested method aids in the detection of plant illness and offers treatments that may be utilized to fend off the disease. Each plant species has been recognized and renamed in order to build an appropriate database. This test-database is then used to check for correctness and confidence in the project. After that, we'll train our classifier with real-world data, and the result will be accurately predicted. For prediction, we utilize a Convolution Neural Network (CNN), which consists of several layers. Additionally, a drone model prototype is being developed. It can be used for live broadcast of large-scale farmland. It is connected to a high-resolution camera to take photos of plants. As an input to the software, the software tells us whether the plants are healthy according to the photos. Using our code and training model, we have reached an accuracy rate of 78%. Our software provides us the name of the plant species and its confidence level as well as remedial measures that can be used as remedial measures.