Calibrating laser scanner data from snow surfaces: Correction of intensity effects

https://doi.org/10.1016/j.coldregions.2015.10.005Get rights and content

Abstract

Terrestrial laser scanning data have become more and more commonly used in cryospheric studies as the commercial instruments are getting cheaper and more user-friendly. We have studied the usability of laser scanning intensity data in remote sensing of snow-covered surfaces by focusing on two topics: the effect of incidence angle on the intensity data and the depth which the backscattered laser beam represents. The measurements were made with a phase-based laser scanner using 650–690 nm wavelength. For some of the snow backscatter vs. depth studies measurements were also made with a pulse-based scanner at 905 nm. The incidence angle effect was studied by rotating a snow surface sample relative to the scanner and measuring the difference in the intensity values. The experiment was repeated for different snow types. The snow pack layer that the backscattered laser signal represents was studied by inserting black metal plates horizontally into the snow pack and measuring the changes in the intensity values with plates at different depths. The results suggest that the snow type has no effect on the incidence angle effect and that for dry snow the backscattering of the laser beam takes place from the very surface, but for wet snow, the majority of the signal is backscattered from 0.5 to 1 cm depth. An empirical correction function for the incidence angle effect is also presented.

Introduction

As the climate is changing rapidly, it is crucial to understand the processes that affect the climate systems. The behavior of snow and ice cover in the changing climate is still fairly poorly known and that introduces considerable uncertainty to the climate models. Better understanding is required on the links between snow geophysical and scattering properties, changing snow and ice-covered area, timing of snow melt, and the surface albedo, which is an essential climate variable (ECV) defined in the Implementation Plan for the Global Observing System for Climate in Support of the United Nations Framework Convention on Climate Change (UNFCCC) (http://unfccc.int/2860.php). Tilt effects are also crucial because they affect the measurement of snow and glacier albedo. Therefore, knowing the effect of incidence angle of the incoming radiation to the snow/ice surface is important in these applications. The measurement geometry effects and their correction have been recently studied by Sicart et al. (2001)) and Weiser et al. (2015). The improvement of the global climate models requires data sets that cover large areas. In practice, this means satellite products. The validation of these products would benefit from in situ data that covers large areas. However, these data are typically not available. Terrestrial laser scanning (TLS) and mobile laser scanning show great potential for gathering the data for validation as well as for analysis (e.g., Egli et al., 2012, Kenner et al., 2011, Kukko et al., 2013).

Terrestrial laser scanning applications on snow surfaces have concentrated on the use of range data (Arnold et al., 2006, Helfricht et al., 2014, Hood and Hayashi, 2010, Prokop, 2008, Prokop et al., 2008, Várnai and Cahalan, 2007). It has proven to be a useful method for mapping inaccessible and dangerous areas, such as potential avalanche sites (for example, Schaffhauser et al., 2008). Some first attempts have been made to use the airborne laser scanning intensity for glacier surfaces (Höfle et al., 2007, Lutz et al., 2003), but the TLS intensity data have, to the best of our knowledge, not yet been widely applied for snow-covered surfaces. The laser reflection and multiple scattering in the snow layer also plays an important role in the usability of TLS range data from snow surfaces. Prokop (2008) found limitations in the application of TLS range measurements for operational avalanche forecasting: if the snow surface was wet and the snow grain size was large (N1 mm), only 50% of the emitted laser signal was received, depending on the angle of incidence. To improve the applicability of TLS data, more accurate knowledge on the incidence angle effect is important. It is also important to know how much the diffuse reflection of laser entering the snowpack contributes to the backscattered signal being detected (cf. Kaasalainen et al., 2006). This is particularly important in mobile laser scanner applications, which typically span a larger area than stationary TLS, resulting in a great variety of incidence angles and point densities, which, in turn, affect the accuracy of results.

Calibrated laser scanner intensity data can be used in segmentation and classification of the range data (Höfle et al., 2007, Yan et al., 2012). Radiometric calibration systems have been developed to enhance the comparability of different scans (Ahokas et al., 2006, Coren and Sterzai, 2006, Höfle et al., 2007, Kaasalainen et al., 2009, Wagner et al., 2006). In our previous studies on the applicability of laser scanning on snow surface mapping (Anttila et al., 2011, Kukko et al., 2013), we have found some challenges related to the measurement and calibration of TLS data: one of these is the considerable effect the incidence angle has on the intensity data (e.g., Anttila et al., 2011). We have found that the object backscattering properties affect the incidence angle effect (Kaasalainen et al., 2011, Krooks et al., 2013). The backscattering properties of snow depend on the grain size and shape (Kaasalainen et al., 2006) and the surface structure (Zhuravleva and Kokhanovsky, 2011). However, our previous studies have suggested that the snow type may have no effect on the incidence angle effect (Anttila et al., 2011). In this paper, we address this issue by studying the mean intensity value of several different snow surfaces in different incidence angles.

To relate the snow intensity parameters to grain properties, it is important to know how much the laser signal penetrates into the snowpack, and, most importantly, which part of the snow layer the backscattered signal represents. Prokop (2008) has studied this using the range data by placing reflective foils and blankets on the snow and comparing the range data of the different surfaces. He found that there was a less than 1 cm difference in the surface height values. We have studied the same topic using the intensity values. We placed matt black painted metal plates horizontally in the snow pack at different depths and measured the changes in the intensity value. We repeated the measurements with different snow types.

The studies presented in this paper are made on taiga snow in the boreal forest zone. After the introduction, in Section 2, we present the methodologies used for both the experiments. In Section 3, we present the results, and in Section 4, we discuss the results in more detail.

Section snippets

Methods

The measurements used in this study were made in Kirkkonummi, Southern Finland (60.1°N, 24.5°E). The snow in Southern Finland is typical taiga snow with ice lenses and various layers of different density and crystal type (Sturm et al., 1995). The relevant geophysical properties of each measured snowpack, including crystal types and sizes, layer structure of the snow pack and surface roughness, were documented during the measurements. In addition to these, overall weather conditions, such as air

Incidence angle

The mean normalized intensities as a function of incidence angle for different samples are shown in Fig. 4. The incidence angle dependency of intensity seems to be similar for all the measured snow types. The error bars in Fig. 4 present the standard deviation of the intensity values. The errors grow together with the incidence angle. This is likely to be caused by the smaller sample sizes of the larger incidence angles. However, the intensity values were approximately normally distributed with

Incidence angle effect

The results suggest that the incidence angle effect is similar for most snow types and wavelengths measured here and therefore the same correction function could be used. The incidence angle effect measurements with LeicaHDS6100 also included a sample of wet snow. There were some problems with errors in the intensity values between scans with this sample. We could not find an obvious reason for this, so the data were excluded from the analysis. However, it is noteworthy that despite this, when

Conclusions

The results presented here indicate that

  • The incidence angle dependence of laser scanner intensity is similar for different snow types. This indicates a prospect of correcting the incidence angle effect with a single function for many types of snow. We also experimented an empirical form cos α of a correction function and found it to be practical in this case.

  • The incidence angle dependence of laser scanning intensity appears similar for all wavelengths between 500 and 1000 nm. More measurements

Acknowledgments

This research was funded by the Academy of Finland research projects “New Techniques in Active Remote Sensing: Hyperspectral laser in Environmental Change Detection” and “Mobile Hyperspectral Laser Remote Sensing” (grant numbers 218144 and 137925). The authors would like to thank Matti Lehtomäki from FGI and Terhikki Manninen from the Finnish Meteorological Institute for all their help and support during the measurements and writing the paper.

References (30)

  • N.S. Arnold et al.

    Evaluating the potential of high-resolution airborne LiDAR data in glaciology

    Int. J. Remote Sens.

    (2006)
  • F. Coren et al.

    Radiometric correction in laser scanning

    Int. J. Remote Sens.

    (2006)
  • L. Egli et al.

    Dynamics of snow ablation in a small Alpine catchment observed by repeated terrestrial laser scans

    Hydrol. Process.

    (2012)
  • C. Fierz et al.

    The International Classification for Seasonal Snow on the Ground

    IHP-VII Technical Documents in Hydrology N°83, IACS Contribution N°1

    (2009)
  • T. Hakala et al.

    Full waveform hyperspectral LiDAR for terrestrial laser scanning

    Opt. Express

    (2012)
  • Cited by (12)

    • An intensity-enhanced method for handling mobile laser scanning point clouds

      2022, International Journal of Applied Earth Observation and Geoinformation
      Citation Excerpt :

      In this section, we introduce the related works of intensity enhancement on the two aspects, i.e. intensity correction and intensity denoising in point clouds. Many previous researches on intensity correction of point-cloud data mainly focused on analyzing the relationship between the measured intensity and the various factors, i.e., temperature (Thundathil et al., 2021), humidity (Anttila et al., 2016), geometric characteristics (Höfle and Pfeifer, 2007; Luo et al., 2019), etc. Once the relationship is determined, the mathematical model is built to implement the task of intensity correction in point clouds.

    • Three dimensional mapping of forest canopy equivalent water thickness using dual-wavelength terrestrial laser scanning

      2019, Agricultural and Forest Meteorology
      Citation Excerpt :

      Calibration is needed for numerous factors that affect the TLS intensity data, including the instrumental effects, the effects of the target distance, and the effects of the incidence angle of the laser beam. Such effects have been highlighted in numerous studies, and methods to calibrate the intensity to apparent reflectance have been successfully developed for various TLS instruments (Anttila et al., 2016; Blaskow and Schneider, 2014; Elsherif et al., 2018; Höfle and Pfeifer, 2007; Jutzi and Gross, 2009; Kaasalainen et al., 2011; Krooks et al., 2013; Tan and Cheng, 2016; Zhu et al., 2017). The calibrated TLS intensity data can then be used to estimate EWT in 3D, using a single shortwave infrared laser wavelength (Zhu et al., 2015; Zhu et al., 2017) or a Normalized Difference Index (NDI) of two laser wavelengths (Elsherif et al., 2018; Gaulton et al., 2013; Junttila et al., 2018, 2016).

    • What Future for Mountain Glaciers? Insights and Implications From Long-Term Monitoring in the Austrian Alps

      2016, Developments in Earth Surface Processes
      Citation Excerpt :

      In the last decade, LiDAR (Baltsavias, 1999; Wehr and Lohr, 1999) systems have been used to observe snow depths, using Terrestrial Laser Scanning (TLS) on individual slopes and Airborne Laser Scanning (ALS) over entire mountain catchments. Anttila et al. (2015) showed that the laser signal is backscattered from a depth of 0.5–1 cm for wet snow, and from the very surface for dry snow conditions. This low penetration depth allows quantification of snow depth and snow depth changes.

    • Laser Scanner in Construction Quality Control of Steel Industrial Buildings

      2023, Practice Periodical on Structural Design and Construction
    • Multispectral terrestrial laser scanning: New developments and applications

      2018, International Geoscience and Remote Sensing Symposium (IGARSS)
    View all citing articles on Scopus
    1

    Former Finnish Geodetic Institute.

    View full text