Enhancing Navigation Benchmarking and Perception Data Generation for Row-based Crops in Simulation

TL;DR

This paper introduces a synthetic dataset and simulation framework to improve navigation and perception in precision agriculture, enabling faster development and evaluation of autonomous crop navigation systems.

Contribution

It provides a novel synthetic dataset and virtual scenarios for training and testing navigation algorithms in agriculture, along with an automatic method to explore various field geometries.

Findings

Deep segmentation networks trained on the dataset show accurate crop perception.

Benchmarking demonstrates improved navigation performance in simulated environments.

The framework accelerates development of autonomous agricultural robots.

Abstract

Service robotics is recently enhancing precision agriculture enabling many automated processes based on efficient autonomous navigation solutions. However, data generation and infield validation campaigns hinder the progress of large-scale autonomous platforms. Simulated environments and deep visual perception are spreading as successful tools to speed up the development of robust navigation with low-cost RGB-D cameras. In this context, the contribution of this work is twofold: a synthetic dataset to train deep semantic segmentation networks together with a collection of virtual scenarios for a fast evaluation of navigation algorithms. Moreover, an automatic parametric approach is developed to explore different field geometries and features. The simulation framework and the dataset have been evaluated by training a deep segmentation network on different crops and benchmarking the resulting navigation.