Differential Analysis of Multispectral Images for Terrain Identification

2026-07-10

AI Summary

This work addresses the challenge of reliable terrain perception under complex conditions such as low illumination, shadows, and ambiguous surface materials, where conventional RGB-based methods often fail. To overcome these limitations, the authors propose DRIFT, a novel framework that fuses raw multispectral bands with illumination-invariant band ratio representations. DRIFT employs a dual-stream residual network to model both feature types and introduces a differential fusion branch to explicitly capture discrepancies between absolute spectral measurements and ratio-derived cues, thereby enhancing robustness against noise and unreliable spectral data. The method effectively mitigates the adverse effects of illumination variations and sensor gain fluctuations, is suitable for deployment on edge devices, and demonstrates significant performance gains over strong baselines on datasets featuring oil-contaminated soil and waterlogged grass, confirming its superiority in scenarios with challenging lighting and thermal disturbances.

Abstract

Reliable terrain understanding is a prerequisite for autonomous robot navigation. Yet, the widespread RGB-based perception can fail under low illumination, shadows, and material ambiguities. In this work we propose DRIFT, a lightweight multispectral framework that combines raw spectral bands and illumination-tolerant band-ratio representations through a dual-stream residual architecture and a differential fusion branch. Band ratios attenuate multiplicative acquisition effects (illumination/sensor gains), while the differential fusion explicitly highlights discrepancies between absolute-band and ratio-derived cues, which improves the robustness to noisy or partially unreliable spectral measurements. In the paper (i) we evaluate DRIFT on a new oil-on-soil multispectral dataset acquired using a MicaSense RedEdge-P camera mounted on an Unmanned Aerial Vehicle, and (ii) we provide an additional controlled study on water-on-grass under varying illumination and thermal perturbations (hot/cold water) to analyze NIR-sensitive effects. DRIFT consistently improves over strong baselines, while remaining compatible with edge deployment.

Problem

• terrain identification
• multispectral images
• autonomous navigation
• illumination robustness
• spectral ambiguity

Innovation

• multispectral imaging
• band ratio
• differential fusion
• illumination invariance
• terrain identification