IoT-based smart trap solution for pest control in sustainable agriculture

Abstract

With the increase in the world population, the demand for food is assuming unprecedented proportions, and ensuring food security (capability to produce food and make it available to the population meeting the minimum nutritional need) is a matter of enormous concern for most countries. The expansion of growing areas is one of the direct consequences of this demand. However, as plantations expand, a suitable environment is promoted to reproducing and establishing undesirable insects due to the abundance of food and sometimes the absence of natural enemies. These insects (known as “pests”) feed on grains, fruits, and leaves, causing plantation degradation and considerable financial losses. As a control method, chemical pesticides are widely used in plantations which chains other problems: poisoning of people and animals, contamination of air, soil, and water in general. Developing environmentally sustainable and viable cost-effective solutions for pest control presents a challenge and an opportunity to ensure desirable food security with quality, healthy and poison-free foods. The use of advanced technology in agriculture has the potential to develop such sustainable control. The combination of advanced technologies such as the Internet of Things (IoT), Cloud Computing, Artificial Intelligence (AI), and Computer Vision (CV) offer proposals for automating and monitoring plantations, storing collected data, performing computational analysis for decision making, and advanced data viewing. The coffee culture is highly relevant in Brazilian agribusiness and, like other cultures, it also faces the pest control challenge. One of the most harmful insects to coffee plantations is known as “Coffee-berry-borer” (CBB) (Hypotenemus hampei), and this dissertation explores the need to control its population by proposing an innovative solution that combines the most advanced technologies and the strict alignment with the concepts of sustainability in the production of quality food. Therefore, throughout the dissertation conception, a review of related literature was carried out to know biology of CBB, its life cycle, and the main approaches to population control, pointing out the respective advantages and disadvantages. Another review was the survey and analysis of insect identification techniques, emphasizing which classes of insects are most suitable in each identification process. Considering the technological requirements, two additional reviews were carried out: The first was about the main IoT concepts highlighting the Agriculture 4.0 (IoT applied in agriculture). And the last review was about the concepts of Computer Vision as a framework to implement insect identification by image analysis. Next, an integrated solution is proposed which combines a smart trap with a location sensor by Global Positioning System (GPS), camera to acquire images of insects, embedded hardware and software for image analysis, and actuators to capture or purge insects. Finally, the smart trap integrates via mobile network into a software layer, known as Middleware, whose function is to receive the data, store it in a database so that the developed Web application, via Internet, can access it and present it to the end-user. The solution was validated firstly to find the optimized value for the binarization threshold within the Computer Vision routine and secondly to ensure the integrated data transmission from the rural environment through the Middleware until the end-user application. Finally, the solution is demonstrated and is ready for use in coffee plantations. Therefore, this study proposes a highly technological solution for sustainable pest control in coffee culture and with the possibility of being used in several other agricultural cultures.