Thermal Imaging-Assisted Infection Classification (BoF) for Brinjal Crop

In the development of economy, agriculture has always played an important role for different nations; since it is considered to be the main source of income, food, and employment to rural populations in the country, owing to diversified geographical locations, environmental conditions, and pest attacks, it is of prime importance to devise technological-assisted methods to monitor and provide early remedial actions for the damage and infections to the crop. Algorithm proposed focuses on health monitoring of brinjal crop using digital thermal imaging. Paper aims to identify the plant disease by analyzing thermal images of brinjal leaves. Infrared images are rich in important hidden details that are not visible due to their low contrast and blurring. Experiment was conducted on two sets of images, first set comprising of healthy and infected thermal images and second comprising of normal RGB capture of healthy and infected images; 30 to 35 images per crop per set were acquired, total dataset analyzed had 1160 images, and the process of identification was implemented via bag of features (BoF), under the umbrella; feature extraction was carried out by SIFT operator, and classification was performed using classification MLSTSVM. Simulation was implemented using MATLAB 2018b. Results showed that duration of the process was less for RGB images by a margin of approximately 6 secs, but the accuracy efficiency achieved was more for thermal images by margin of 3%, having 87% in all. From the results, it can be concluded that however duration required for the identification was more for thermal images but still percentage accuracy is more for thermal images; thus, thermal image-assisted algorithm can be employed for crops in remote scenarios where accuracy plays a vital role.