Philippine Journal of Science
151 (4): 1313-1320, August 2022
ISSN 0031 – 7683
Date Received: 06 Oct 2021

 

Classification and Percent Severity of Pechay Damage
Caused by Cutworm (Spodoptera litura)

Anton Louise P. de Ocampo¹, Angelica D. Manalo², Mary Jane L. Silva²,
April Kristine I. Villanueva², and Romel Brian B. Florendo²*

¹Electronic Systems Research Center
²College of Engineering, Architecture, and Fine Arts (CEAFA)
Batangas State University–Alangilan Campus,
Alangilan, Batangas City 4200 Philippines

*Corresponding author: rbrianflorendo@gmail.com

[Download]
Ocampo AL et al. 2022. Classification and Percent Severity of Pechay Damage Caused by
Cutworm (Spodoptera litura). Philipp J Sci 151(4): 1313–1320. https://doi.org/10.56899/151.04.02

ABSTRACT

Agriculture is one of the most important sectors of any country, especially the Philippines. Pechay is the sixth-ranking crop for agricultural production in Batangas, according to the Philippine Statistics Authority (PSA). However, pechay is constantly threatened by different pests, such as the occurrence of cutworm (Spodoptera litura), which primarily feeds on its leaves, stalks, and stems and even cuts off the plants that destroy the entire leaf in no time at all. Detection of the presence of pests at the earliest can help farmers to employ the necessary intervention to mitigate the spread of the infestation. This paper proposes a system that can classify whether pechay plants are healthy or damaged and assess the severity of the damage using image processing techniques and machine learning. Images of pechay plants are gathered and pre-processed to remove the background, resize, and enhance. Then statistical measures derived from GLCM (gray level co-occurrence matrix) are used with diagnostic feature explorer (DFE) to select the most appropriate features for the classification. The clustering to assess the severity of damage used k-means to segment the damaged areas from the entire leaf. The system can classify healthy and destroyed plants and grade the severity of damage with an accuracy of 88.31%.