Vision Intelligence for Mobile Agro-Robotic System

Abstract

We developed an intelligent vision system for mobile robot field operations. Fuzzy logic was used to classify crops and weeds. A genetic algorithm (GA) was used to optimize and tune fuzzy logic membership rules. Field studies confirmed that our method accurately classified crops and weeds throughout their growth cycle. After separating out weeds, an artificial neural network (ANN) was used to estimate crop height and width. The r 2 for estimating crop height was 0.92 for training data and 0.83 for test data. A geographic information system (GIS) was used to create a crop growth map.