A novel approach for vegetation classification using UAV-based hyperspectral imaging

Tetsuro Ishida,Junichi Kurihara,Fra Angelico Viray,Shielo Baes Namuco,Enrico C Paringit,Gay Jane Perez,Yukihiro Takahashi,Joel Joseph Marciano

doi:10.1016/j.compag.2017.11.027

Abstract

The use of unmanned aerial vehicle (UAV)-based spectral imaging offers considerable advantages in high-resolution remote-sensing applications. However, the number of sensors mountable on a UAV is limited, and selecting the optimal combination of spectral bands is complex but crucial for conventional UAV-based multispectral imaging systems. To overcome these limitations, we adopted a liquid crystal tunable filter (LCTF), which can transmit selected wavelengths without the need to exchange optical filters. For calibration and validation of the LCTF-based hyperspectral imaging system, a field campaign was conducted in the Philippines during March 28–April 3, 2016. In this campaign, UAV-based hyperspectral imaging was performed in several vegetated areas, and the spectral reflectances of 14 different ground objects were measured. Additionally, the machine learning (ML) approach using a support vector machine (SVM) model was applied to the obtained dataset, and a high-resolution classification map was then produced from the aerial hyperspectral images. The results clearly showed that a large amount of misclassification occurred in shaded areas due to the difference in spectral reflectance between sunlit and shaded areas. It was also found that the classification accuracy was drastically improved by training the SVM model with both sunlit and shaded spectral data. As a result, we achieved a classification accuracy of 94.5% in vegetated areas.

Highlights

Spectral reflectance data collected from vegetated areas can provide very valuable information on factors such as the presence or absence of certain tree species, plant growth stages, and plant diseases
We developed a snapshot hyperspectral imaging system that does not require accurate global positioning system (GPS) measurements so that the acquired dataset can be processed by a simple image processing technique, detailed below
Model, we classified the spectral reflectances obtained by the image processing technique described in Section 3, and created a high-resolution classification map of the study area

Summary

Introduction

Spectral reflectance data collected from vegetated areas can provide very valuable information on factors such as the presence or absence of certain tree species, plant growth stages, and plant diseases. Because of the flexibility of spectral bands, hyperspectral imaging with LCTF technology is applicable to a wide variety of remote-sensing applications This technology was first put into practical use on a space-borne instrument by Hokkaido University, and it has already been mounted on several microsatellites developed by Tohoku University and Hokkaido University (e.g., Sakamoto et al, 2015). The UAV-based hyperspectral imaging system described in this study can realize aerial images with a resolution on the order of tens of millimeters It could be useful for leaf-scale plant disease detection when combined with the ML approach. This new survey platform using LCTF technology will make a significant contribution to future precision agriculture research. We present a UAV-based high-resolution vegetation classification map, and evaluate the validity of this new platform

Observations

Data processing

Results and discussion

Conclusion