CNN-Based Terrain Classification with Moisture Content Using RGB-IR Images

Tomoya Goto, Genya Ishigami

Research output: Contribution to journalArticlepeer-review

Abstract

Unmanned mobile robots in rough terrains are a key technology for achieving smart agriculture and smart construction. The mobility performance of robots highly depends on the moisture content of soil, and past few studies have focused on terrain classification using moisture content. In this study, we demonstrate a convolutional neural network-based terrain classifi¬cation method using RGB-infrared (IR) images. The method first classifies soil types and then categorizes the moisture content of the terrain. A three-step image preprocessing for RGB-IR images is also integrated into the method that is applicable to an actual envi-ronment. An experimental study of the terrain classi¬fication confirmed that the proposed method achieved an accuracy of more than 99% in classifying the soil type. Furthermore, the classification accuracy of the moisture content was approximately 69% for pumice and 100% for dark soil. The proposed method can be useful for different scenarios, such as small-scale agri¬culture with mobile robots, smart agriculture for mon¬itoring the moisture content, and earthworks in small areas.

Original languageEnglish
Pages (from-to)1294-1302
Number of pages9
JournalJournal of Robotics and Mechatronics
Volume33
Issue number6
DOIs
Publication statusPublished - 2021 Dec

Keywords

  • CNN
  • Ma¬chine learning
  • Moisture content
  • Terrain classification

ASJC Scopus subject areas

  • Computer Science(all)
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'CNN-Based Terrain Classification with Moisture Content Using RGB-IR Images'. Together they form a unique fingerprint.

Cite this