Extraction of Pest Insect Characteristics Present in a Mirasol Pepper (Capsicum annuum L.) Crop by Digital Image Processing

Mireya Moreno-Lucio,Carlos Alberto Olvera-Olvera,Héctor A. Guerrero-Osuna,Luis F. Luque-Vega,Antonio Del Rio-De Santiago,Luis Octavio Solís-Sánchez,Gustavo Espinoza-García,Celina Lizeth Castañeda-Miranda,Ma. del Rosario Martínez-Blanco

doi:10.3390/app112311166

Mireya Moreno-Lucio, Carlos Alberto Olvera-Olvera + Show 7 more

Open Access

https://doi.org/10.3390/app112311166

Copy DOI

Abstract

One of the main problems in crops is the presence of pests. Traditionally, sticky yellow traps are used to detect pest insects, and they are then analyzed by a specialist to identify the pest insects present in the crop. To facilitate the identification, classification, and counting of these insects, it is possible to use digital image processing (DIP). This study aims to demonstrate that DIP is useful for extracting invariant characteristics of psyllids (Bactericera cockerelli), thrips (Thrips tabaci), whiteflies (Bemisia tabaci), potato flea beetles (Epitrix cucumeris), pepper weevils (Anthonomus eugenii), and aphids (Myzus persicae). The characteristics (e.g., area, eccentricity, and solidity) help classify insects. DIP includes a first stage that consists of improving the image by changing the levels of color intensity, applying morphological filters, and detecting objects of interest, and a second stage that consists of applying a transformation of invariant scales to extract characteristics of insects, independently of size or orientation. The results were compared with the data obtained from an entomologist, reaching up to 90% precision for the classification of these insects.

Highlights

Agriculture 4.0 replaces traditional production methods and global agricultural strategies by integrating them into a value chain [1], allowing the agronomic industry to seek the interconnection of the systems available for the field and adaptability of production systems, improving the rotation of crops to achieve a higher level of production and efficiency of production systems by optimizing the efficient use of water, fertilizers and phytosanitary products
The digital processing of sticky yellow trap images has been shown to help detect, classify and count different types of pest insects present in crops [11,19,20,21], so there is a need to apply this method in agricultural areas to help entomologists confirm their detections or classifications
The Scale-Invariant Feature Transform (SIFT) algorithm detected the five genera in images of 17 insects of the species reported in this work with up to 94% accuracy, compared to the image with 5 insects of the species reported, in which an average of 100% accuracy was obtained, suggesting that more insects present in an image increases the accuracy of detection

Summary

Introduction

Agriculture 4.0 replaces traditional production methods and global agricultural strategies by integrating them into a value chain [1], allowing the agronomic industry to seek the interconnection of the systems available for the field and adaptability of production systems, improving the rotation of crops to achieve a higher level of production and efficiency of production systems by optimizing the efficient use of water, fertilizers and phytosanitary products. The presence of pests in open-air crops and under greenhouse conditions is one of the biggest problems faced by producers at a national and international level since it can cause losses of up to 40% of the sown crop [5]

Objectives

Methods

Results

Discussion

Conclusion