Design of an IoT-Based Quantity Controlled Pesticide Sprayer Using Plant Identification

Abstract

This chapter discusses the design of a quantitative controlled pesticide sprayer and the development of an efficient algorithm for plant identification. The whole system is controlled using the raspberry pi and convolutional neural networks (CNN) algorithm for training the proposed model. Once the algorithm identifies the plant by processing the image, it is captured by using a pi camera, and it determines the pesticide and its dosage. The sensors will collect the information related to the plant condition such as humidity and surrounding temperature, which is simultaneously sent to the farmers/agriculture officers through the internet of things (IoT), for the purpose of live analysis, and they are stored using cloud services, making the system suitable for remote farming. The proposed algorithm is trained mainly for three types of plant leaves, which include tomato, brinjal, and chilly. The CNN algorithm scores accuracy of 97.2% with sensitivity and specificity of 0.94 and 0.95, respectively. The robot is intended to encourage the agriculturists for next-level farming to facilitate their work.

Key Terms in this Chapter

Internet of Things (IoT): Internet of things is the interconnection of computing devices embedded in any objects, enabling them to send and receive data.

Convolutional Neural Networks (CNN): It is a class of deep learning technology mainly used for accurate classification of images.

Raspberry Pi: It is a credit card sized computer especially used for exploring the applications of internet of things (IoT).

Pesticide: The chemicals used to control or kill the pests.

Next-Level Farming: It is defined as providing over-all solutions in other words complete modernization from the start till the end of the agricultural applications.

Remote Farming: It is defined as monitoring and managing of farm remotely through the application of Internet, with less/ without direct human intervention.