Deep Learning Solutions for Pest Detection

Abstract

In today’s world, farmland is more than just a feeding source for billions of people. The Indian economy is highly dependent on agricultural productivity. Since global cultivable land is shrinking, increasing the productivity of existing agricultural land is more important. As a result of this need, the scientific community has focused its efforts on creating efficient and long-term methods for increasing crop production. Plant diseases and pests are the most significant factor that reduces farm yield and its quality. Therefore, in agronomy, the timely detection of plant diseases and pathogens has an important role; it affects the crop yields and the farmer’s income and gradually increases the cost of production over time. Traditional methods use image-processing techniques to classify plant pathogens, pests. Deep learning has now generated advances in digital image processing that far outperform the conventional approaches. This article introduces object detection, the advances in modern-day agriculture, and the investigation of intrinsic and extrinsic factors influencing the automatic detection of plant pathogens, using deep learning. This article also explores a preliminary study to develop a novel approach for fast, accurate, and robust rice plant pest detection.