Plant detection using de-segmentation and color constancy algorithms for robotic navigation

• Grant number: 20/12710-6
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): March 01, 2021
• Effective date (End): February 28, 2022
• Field of knowledge: Engineering - Mechanical Engineering
• Principal Investigator: Marcelo Becker
• Grantee: Gabriel Corrêa de Oliveira
• Host Institution: Escola de Engenharia de São Carlos (EESC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

With the growing demand for food due to world population growth, technology has deeply permeated the agricultural sector in recent decades. Within this context, robotics was one of the areas in which there was the greatest investment and technological development. The purposes of robots in the field are numerous: cost reduction, greater productivity and efficiency, greater use of area and resources, among others. The Mobile Robotics Lab of EESC-USP operates in this field through several robotic platforms, among them, the TerraSentia mobile robot will be used in this research. The objective of this robot is to collect data on plant characteristics at low cost and quickly, accurately and, mainly, autonomously. Due to this last point, many challenges arise in order to achieve the complete autonomy of the robot, among which we can mention the presence of numerous and dense foliage that protrudes on the trail where TerraSentia needs to move, making it difficult to orient. That said, this undergraduate research project will study and apply computer vision methods, more specifically segmentation and color constancy, to remove the pixels corresponding to the leaves that hinder TerraSentia's autonomous navigation, impacting the performance of the entire algorithm.