TinyML-Based Human and Animal Movement Detection in Agriculture Fields in India

Tiny machine learning (TinyML) is blooming in ML field that deals with the performance of machine learning on extremely limited edge devices. Recently, a variety of data-intensive and time-sensitive Internet of Things (IoT) applications are using deep learning algorithms more often. As a result, various fresh strategies, such as deep neural networks (DNN) model deployment on MCUs, have become challenging task as they lack resources like memory. But recent advancement in the field of TinyML promises to open up a brand-new category of edge applications. In this paper, animal detection in farmlands is addressed using TinyML models deployed on SparkFun Edge device which supports high-resolution tiny camera attached to the device board itself in view of protecting farmlands which are widely being developed in India from animal attacks since TinyML provides the path for the creation of unique apps and services that do not require the cloud's ubiquitous computing support, which consumes power and poses dangers to data security and privacy. Also the models are tested on Google Colab. The results are obtained and discussed in result section.