Detection of invasive plants in sugarcane crop using hyperspectral imagery and fuzzy logic

Abstract

The constant increase in fuel consumption associated with the search for renewable energy sources has led to the growing of sugar cane, the raw material for ethanol production, with a demand in volume and technology development, especially in the state of Sao Paulo, where the majority of sugar cane production is located. As a result of this positioning in the energy sector, various research institutions have been working to develop mechanisms that will reduce costs and improve the different stages of the production of ethanol, production, ensuring the increased quality of the product and mitigation of impacts to the environment, from the perspective of sustainability. Although the use of hyperspectral sensor and processing techniques of digital images in real time are not yet applied to crops in Brazil, studies have shown great potential to be applicable in agricultural machinery, as a tool to support decision making and sustainable management of production. This project falls within the framework of Precision Agriculture, and aims at developing a tool for detection of invasive plants through a system to capture hyperspectral imaging and fuzzy logic and provide basis for further development of equipment that allows variable-rate application of herbicides, seeking the management of sugarcane in an beneficial cost-benefit ratio. By the end of the project it is expected to have a tested, calibrated and validated prototype of an acquisition system and image processing software with hyperspectral classification using fuzzy logic for the detection of invasive plants in cultivation of sugar cane, deepening the understanding of and allowing application of hyperspectral imaging systems for the advancement of Precision Agriculture in Brazil. Additionally, it will produce a basis for the development of a variable rate application system of herbicides, and will warrant further studies using hyperspectral imagery in Brazil, contributing to the implementation and technology transfer to the productive sector. At the same time, the project will be a major progress in the area of precision agriculture in production chains, contributing to the acquisition of skills and infrastructure to further continue research in the topic at the College of Agricultural Engineering. (AU)