Vol. 10, Special Issue 10 (2021)

Author(s):
R Prabha, JS Kennedy, G Vanitha and N Sathiah

Abstract:
Fall armyworm has become a major concern for maize farmers in recent years, as it has resulted in significant yield losses in the maize field. A fall armyworm infestation might be detected automatically using a machine learning system, allowing for faster and accurate scouting of farmers' field operations. However, it is tedious for creating a machine-learning algorithm to discern between the target fall armyworm infestation and other sources of weeds, soil in a typical field. So, a vast amount of human-generated training data is required to train a machine learning system to consistently detect a specific fall armyworm infestation in the maize field. In this study, we created an image repository for different severity levels of fall armyworm infestation in maize. All of the high-quality photographs were shot with a digital camera against a variety of backgrounds with distinct light intensities in different locations. Visual scale ratings were also given to fall armyworm infestation in maize.