Pests in the Coconut Cultivation Industry: An IT Based Solution for Pest Detection in the Coconut Cultivation Industry

Abstract

Coconut cultivation industry plays a pivotal role in Sri Lanka. It involves many industries such as desiccated coconut, coconut water and coconut oil etc. So many peoples’ lives depend on this cultivation. Although year by year, we can see a decline in coconut production. Main reason for thisunfortunate state is mainly pests. If coconut cultivator is able to identify relevant pest in early stage, it is easy to prevent spreading before it becomes worsen by spreading to the whole coconut plantation. But the main challenge is the identification of pests correctly and applying suitable remedies for them. Sometimes it is hard for coconut cultivated farmers to contact and get instructions from relevant officers to eradicate pests that arise in coconut cultivation because of telecommunication problems and those officers also have to cover a big area. It is not an easy job for them as well. Cultivators cannot identify the specific pest, so they cannot make appropriate remedy and some of them do not aware of it in any means. Due to the covid19 situation, this problem became worsened. To overcome these problems and challenges, decided to introduce an IT-based solution to identify pests with relevant remedysolutions.The selected algorithm wasMask RCNN for the development oftheproject. Dataset was collected by available images in the internet. There were three pest categories of images with 152 images from each category. Those images were then subjected to a bounding box and mask labeling method and some pre-processing steps. This developed pest detection system is capable of identifying three most harmful pests present in Sri Lanka with high accuracy. Another special feature is, it has the capability of identifying many pests present in any image with a uniform or complex background. And also, one of the pests among the three selected pests is very small, so thesystem is capable of identifying thatsmall pest accurately as well. Mostimportantly the system is giving remedy solutions for the identified pest. Those features represent the research gap that was absent in many existing systems. The outcome of this project will help coconut cultivators and who are interested in coconut cultivation to enhance their productivity by minimizing pests and by making them aware of those problems and relevant solutions.