Spatial Co-Registration of Ultra-High Resolution Visible, Multispectral and Thermal Images Acquired with a Micro-UAV over Antarctic Moss Beds

Abstract

Abstract:In recent times, the use of Unmanned Aerial Vehicles (UAVs) as tools for environmental remote sensing has become more commonplace. Compared to traditional airborne remote sensing, UAVs can provide finer spatial resolution data (up to 1 cm/pixel) and higher temporal resolution data. For the purposes of vegetation monitoring, the use of multiple sensors such as near infrared and thermal infrared cameras are of benefit. Collecting data with multiple sensors, however, requires an accurate spatial co-registration of the various UAV image datasets. In this study, we used an Oktokopter UAV to investigate the physiological state of Antarctic moss ecosystems using three sensors: (i) a visible camera (1 cm/pixel), (ii) a 6 band multispectral camera (3 cm/pixel), and (iii) a thermal infrared camera (10 cm/pixel). Imagery from each sensor was geo-referenced and mosaicked with a combination of commercially available software and our own algorithms based on the Scale Invariant Feature Transform (SIFT). The validation of the mosaic's spatial co-registration revealed a mean root mean squared error (RMSE) of 1.78 pixels. A thematic map of moss health, derived from the multispectral mosaic using a Modified Triangular Vegetation Index (MTVI2), and an indicative map of moss surface temperature were then combined to demonstrate sufficient accuracy of our co-registration methodology for UAV-based monitoring of Antarctic moss beds.

Estimation of evapotranspiration of temperate grassland based on high-resolution thermal and visible range imagery from unmanned aerial systems

“…The advent of low-cost UAS platforms and concomitantly lightweight camera systems in the visible, near-infrared (NIR), and thermal spectral range has motivated their increased use in the remote-sensing community, e.g. precision agriculture applications (Berni et al 2009; Brosy et al 2017; Zhang and Kovacs 2012; Candiago et al 2015; Reineman et al 2013; Link, Senner, and Claupein 2013; Lelong et al 2008; Turner et al 2014; Stefano et al 2017; Vázquez-Tarrío et al 2017). However, studies using UAS-based TIR sensors to map LST and subsequently derive surface turbulent heat fluxes are still rare (Hoffmann et al 2016b; Ortega-Farías et al 2017; Ortega-Farías et al 2016; Brenner et al 2017).…”

Optimizing the Processing of UAV-Based Thermal Imagery

“…We propose several improvements of the basic framework introduced by Turner et al [33], which allows aligning and processing of complex and more challenging datasets to obtaining thermal orthophotos corrected for changing conditions over flight measurements. This improvement makes use of the information from a RGB (or NIR) camera on board the UAV, since in the large majority of the flights, these are taken along to be able to interpret the thermal information.…”

“…Turner et al [33] provided a general framework for processing thermal imagery with UAV data, consisting of three steps:…”

Improved discrimination among similar agricultural plots using red-and-green-based pseudo-colour imaging

“…Each pixel represented a 20Í20 m square (Tadjudin and Landgrebe, 1996). Among the 220 images, six were selected to prepare a dataset like those acquired by commercially available multispectral sensors (Turner et al, 2014). The selected bands were blue (wavelength 479 nm, band 9), green (557 nm, band 17), red (657 nm, band 27), near-infrared (822 nm, band 46), mid-infrared 1 (1.65 µm, band 133), and midinfrared 2 (2.20 µm, band 190).…”