Time of day impact on mapping agricultural subsurface drainage systems with UAV thermal infrared imagery

https://doi.org/10.1016/j.agwat.2021.107071Get rights and content

Highlights

  • UAV obtained thermal infrared (TIR) imagery can map farm field drainage pipes.

  • The TIR drainage pipe response can change considerably from sunrise to sunset.

  • High relative humidity can adversely impact TIR drainage mapping image quality.

  • Late morning to late afternoon is best for UAV TIR drainage mapping.

  • UAV TIR drainage mapping at sunrise or sunset can often give excellent results.

Abstract

Due to economic and environmental considerations, there exists a need for effective, efficient, and nondestructive methods for locating buried agricultural drainage pipes. Previous research indicates that thermal infrared (TIR) imagery obtained with an unmanned aerial vehicle (UAV) has potential for mapping agricultural subsurface drainage systems, thereby warranting further investigation to determine the best time of day to conduct these UAV TIR surveys. Accordingly, a set of sunrise to sunset UAV TIR surveys were carried out at four different farm field sites in Ohio, U.S.A. Late morning through late afternoon UAV TIR surveys were generally found to work well for determining drainage system patterns. During late morning through late afternoon, the apparent radiant temperature of the soil surface over the drain lines was higher than between the drain lines (i.e., emitted TIR radiation from the soil surface over a drain line was greater than between the drain lines). Conversely, near sunrise or sunset, the UAV surveys often showed the apparent radiant temperature of the soil surface over the drain lines to be lower than between the drain lines (i.e., less emitted TIR radiation over the drain lines than between drain lines). Some excellent UAV TIR drainage mapping results were obtained near sunrise/sunset due to TIR drain line responses that were more easily distinguished from those of farm field operations. However, difficulties were occasionally encountered processing this sunrise/sunset TIR imagery, likely due to the impact on image quality from high relative humidity during these times of the day. Consequently, strictly on a consistency of success basis alone, late morning through late afternoon are the best times for locating drainage pipes with UAV TIR surveys; however, in certain cases, UAV TIR surveys at sunrise/sunset can provide exceptional drainage pattern recognition. These results provide valuable guidance for those considering UAV TIR drainage mapping surveys.

Introduction

The widespread adoption of subsurface drainage practices to remove excess soil water has enabled the Midwest U.S.A. to become one of the most productive agricultural regions in the world. A 1985 economic survey showed that several states within this region (Illinois, Indiana, Iowa, Ohio, Minnesota, Michigan, Missouri, and Wisconsin) had by that year approximately 12.5 million ha containing subsurface drainage systems, with cropland accounting for the vast majority of areas having buried drainage pipes (Pavelis, 1987). These systems are typically comprised of 10 cm diameter older clay tile or newer corrugated plastic tubing drain line networks buried 0.6–1.3 m beneath the surface. Since 1985, a substantial amount of additional agricultural drainage pipe has been installed. Farmers within this region, and other parts of the world, where agricultural drainage practices are common, often need to repair drain lines that are not functioning correctly, or in order to increase crop yields, install new drain lines between the old ones to improve soil water removal efficiency. Either for system repairs or efficiency improvements, locations of the preexisting drain lines are required; however, in most cases, a map of the original subsurface drainage system installation is no longer available. Furthermore, subsurface drainage practices can release substantial amounts of nitrate (NO3-) and phosphate (PO43-) from farm fields into adjacent waterways (Sims et al., 1998, Zucker and Brown, 1998), which in turn degrades surface water bodies on local, regional, and national scales. Risk assessment of this environmental hazard from a farm field perspective calls for knowledge of the installed drainage pipe network, including extent of coverage and drain line spacing distance.

Regardless of whether the need is economic or environmental, finding drain lines with a hand-held tile probe is time-consuming, extremely tedious, and if not careful, can damage buried pipes (Fig. 1a). Using heavy trenching equipment is generally effective, but always causes considerable pipe damage requiring costly repairs (Fig. 1b). Subsurface drainage system patterns can be complex (Fig. 2), further hampering efforts to map drain lines using traditional tile probe or trenching detection methods. Consequently, there is a crucial necessity for effective, efficient, and nondestructive drainage pipe mapping methods. Proximal soil sensing or imagery obtained with unmanned aerial vehicles (UAVs) can potentially provide a viable means for mapping these agricultural subsurface drainage systems.

Ground penetrating radar (GPR), a proximal soil sensing method, has proven capable in many instances of locating buried drainage pipes under a range of soil conditions (Allred et al., 2004, Allred et al., 2005, Allred et al., 2018a, Allred and Redman, 2010, Allred, 2013, Chow and Rees, 1989, Koganti et al., 2020). However, GPR is somewhat inefficient for providing detailed maps of subsurface drainage systems in large farm fields. Both visible (VIS - i.e., wavelengths within the 400–700 nm visible light portion of electromagnetic spectrum) and color infrared (CIR - comprised of green, red, and near infrared wavelengths) airplane photography obtained outside the growing season have been tested to a limited extent for mapping subsurface drainage systems (Naz et al., 2009, Verma et al., 1996). These studies noted that optimal drainage mapping field conditions exist several days after a rainfall event equal or greater than 2.5 or 5.0 cm; although, scheduling an airplane to obtain aerial imagery when these field conditions are present can be problematic and expensive. Alternatively, scheduling an aerial survey with an Unmanned Aerial Vehicle (UAV) is far more flexible and could solve this timing availability issue, but UAVs have not been extensively tested for drainage pipe mapping.

Even though some aerial VIS and CIR drainage pipe detection studies have been conducted, there has been very little research on the use of thermal infrared (TIR) imagery to map drainage pipes. Abdel-Hadi et al. (1970) provides several excellent examples in which TIR imagery detected subsurface features such as pipelines and buried stream channels, thereby indicating that TIR aerial surveys may have the capability for drainage pipe mapping. A laboratory experiment conducted by Woo et al. (2019) demonstrated the potential of TIR imagery for locating buried drainage pipe, further implying that timing relative to rainfall can impact the success of using TIR imagery for this purpose. In a preliminary case study by Allred et al. (2018b), a UAV deploying a TIR camera detected roughly 60% of the subsurface drainage infrastructure known to be present at an agricultural field near Mount Gilead, Ohio, U.S.A. At the same field site used by Allred et al. (2018b), Freeland et al. (2019) found that pairing of Real-Time Kinematic (RTK) Global Navigation Satellite System (GNSS) technology with the UAV TIR survey was essential for accurately locating buried drain lines. Williamson et al. (2019) utilized both UAV multispectral (MS – often including separate narrow bands of blue, green, red, red edge and near infrared wavelengths) and TIR imagery in a complementary approach to map random drain lines in two conservation tillage farm fields near Harlan, Indiana, U.S.A. Tlapáková et al. (2017) documented success using manned aircraft and UAVs with VIS, CIR, near infrared (NIR), and TIR sensors to map agricultural drainage systems in the Czech Republic. Kratt et al. (2020) carried out UAV surveys at two Illinois farm fields four days after a small rainfall event and found visible-color (VIS-C) imagery better than TIR imagery at finding drainage pipe locations. Aerial surveys using a UAV with visible-color (VIS-C), MS, and TIR cameras were conducted by Allred et al. (2020) at 29 agricultural field sites in the Midwest U.S.A., and overall results showed that VIS-C, MS, and TIR imagery detected at least some of the drainage pipe present at 48%, 59%, and 69% of the sites, respectively. Based on this past research, additional investigation is certainly warranted, particularly on the use of UAV TIR imagery for mapping subsurface drainage.

The soil surface directly over top of a drain line is often drier than the soil surface between drain lines. This phenomenon is especially true in the days after a significant rainfall, because the soil over a drain line is dewatered at a faster rate than the soil between drain lines (Smedema et al., 2004). The diurnal cycle in solar radiation received at the ground surface, in conjunction with the thermal conductivity and specific heat capacity of soil pore water being significantly different from that of air within the pores, can result in the drier soil surface over a drain line having a different true kinetic temperature than a wetter soil surface between drain lines (Hillel, 1980, Jenson, 2007).

As indicated by Eq. (1), the Stefan-Boltzmann and Kirchhoff’s laws stipulate that the TIR radiation emitted from an object, Mr (W m−2), and the apparent radiant temperature, Trad (K), are a function of that object’s true kinetic temperature, Tkin (K), and dimensionless emissivity, ε, (Jenson, 2007, Kuenzer and Dech, 2013);Mr=σTrad4=εσTkin4,where the Stefan-Boltzmann constant, σ, equals 5.6695 × 10−8 W m−2 K−4. Soil surface emissivity is impacted by its moisture level, and generally, a wet soil surface will have a higher ε value than a dry soil surface (Jenson, 2007, Mira et al., 2007). A TIR camera can detect soil surface variations in Mr (i.e., and Trad) caused by variations in Tkin and/or ε. Several studies have showed that linear soil surface features depicted in UAV TIR imagery, which exhibited Trad (i.e., and Mr) different than their surroundings, often revealed the presence buried agricultural drain lines (Allred et al., 2018b, Allred et al., 2020, Freeland et al., 2019, Williamson et al., 2019).

The diurnal Tkin range is much greater for a dry soil surface than a wet soil surface, again due to the thermal conductivity and specific heat capacity of soil pore water being significantly different from that of air within the pores (Hillel, 1980, Jackson et al., 1974, Jenson, 2007, Mengistu et al., 2018). From late morning to late afternoon, the Tkin of a dry soil surface is typically higher than for a wet soil surface (Jackson et al., 1974, Jenson, 2007, Mengistu et al., 2018). Abdel-Hadi et al. (1970) conducted experiments with sandy and silty clay soils and found that the emitted TIR radiation for both soils was substantially greater for a dry soil surface than a wet soil surface during daylight from 12:00–18:00. Therefore, from late morning to late afternoon, the soil surface Trad over a drain line is likely to be higher than between drain lines.

Most research indicates that from sunset to sunrise a dry soil surface Tkin will still be slightly warmer than a wet soil surface (Jackson et al., 1974, Jenson, 2007). However, it is possible, during the time from sunset to sunrise, that the emitted TIR radiation from a wetter soil surface to be greater than from a drier soil surface. Based on the Stefan-Boltzmann and Kirchhoff’s laws, this phenomenon could occur due to the wetter soil surface having a higher ε value than the drier soil surface, even when the wetter soil surface has a slightly lower Tkin than a drier soil surface (see Eq. (1)). Experiments carried out by Abdel-Hadi et al. (1970) on sandy and silty clay soils and found the Mr for both soils was somewhat greater for a wet soil surface than a dry soil surface from midnight (0:00) to sunrise (6:00). Therefore, it is possible near sunrise or sunset for the soil surface Trad over a drain line to be less than between drain lines, the reverse of what would occur between late morning to late afternoon.

The common convention for grayscale UAV TIR imagery is for features that have lower Trad to be darker shaded and features that have higher Trad to be lighter shaded. During the time of day from late morning to late afternoon, previous studies (Allred et al., 2018b, Allred et al., 2020, Freeland et al., 2019, Williamson et al., 2019) have shown the drain line response depicted in UAV TIR imagery to be distinct lighter shaded linear features. This late morning to late afternoon TIR drainage pipe mapping result is again most certainly due to drier soil over the drain line having a higher Tkin than wetter soil between drain lines (Jackson et al., 1974, Jenson, 2007, Mengistu et al., 2018). The drainage pipe response obtained by UAV TIR imagery in the early morning hours near sunrise and the evening hours near sunset was unclear prior to this research. Based on previous discussions, it would seem possible during early morning and late evening that the drier soil surface over a drain line could have a Trad (i.e., and Mr) greater, the same, or less than the surrounding soil surfaces between the drain lines. If the Trad at the soil surface over the drain line is the same as the Trad between drain lines, then there would be no TIR drainage pipe response, while a Trad over the drain line less that between drain lines would be depicted as darker shaded linear features within typical UAV TIR grayscale imagery.

Further complicating matters for drainage mapping with UAV TIR surveys conducted in early morning and evening hours is the increased relative humidity that occurs during these times of the day (Ackerman and Knox, 2015, Ahrens, 1988). High relative humidity causes atmospheric absorption and re-emission of TIR radiation (AZoSensors, 2018; FLIR Systems, 2018; Jenson, 2007; Zhang et al., 2016). This high relative humidity atmospheric effect in turn degrades the quality of individual TIR images obtained during a UAV survey, thereby hampering the ability to “stitch” these images together to produce a complete TIR orthomosaic map of the farm field area (PrecisionHawk, 2019, senseFly SA, 2020). Recommendations are that UAV TIR surveys not be carried out if the relative humidity is above 60% (PrecisionHawk, 2019) or 75% (senseFly SA, 2020). Consequently, daytime variations in meteorological conditions, especially relative humidity, can potentially affect TIR image quality, and may need careful consideration when planning a UAV TIR drainage mapping survey.

A large knowledge gap definitely exists on time of day impacts for UAV TIR subsurface drainage mapping. As previously indicated, (1) the TIR drainage pipe response may change during the day, and (2) daytime variations in meteorological conditions can affect TIR image quality. Gaining insight on the best times of day to conduct a UAV TIR drainage mapping survey, and the times of day to avoid, required sets of sunrise to sunset UAV TIR surveys be conducted at four different farm field sites in Ohio, U.S.A., with one of the fields visited twice. Note: When this study was conducted, nighttime commercial UAV flights in the U.S. were not allowed without authorization, and could normally only be carried out between the beginning of morning civil twilight - 20–30 min before sunrise, and the end of evening civil twilight - 20–30 min after sunset (Aitken et al., 2017). Therefore, the overall project goal was to obtain data needed to provide specific guidance on the best time of day, from sunrise to sunset, for UAV TIR drainage mapping, based on (1) the TIR drainage pipe response and (2) meteorological conditions. This project had a stated research hypothesis, “Late morning through late afternoon are the best times to conduct UAV TIR mapping of subsurface drainage systems.”

Section snippets

Materials and methods

Note: This research used much the same equipment and data processing procedures as employed for the Allred et al. (2020) study. Consequently, for the most part, information on equipment and data processing provided in 2.1 UAV survey equipment, 2.2 Data processing of UAV imagery. were previously reported in Allred et al. (2020). The restating (sometimes verbatim) of equipment and data processing information in 2.1 UAV survey equipment, 2.2 Data processing of UAV imagery. is done solely for the

Overall results

Research results are summarized in Table 3. Examples of unmanned aerial vehicle (UAV) thermal infrared (TIR) imagery are shown in Fig. 6, Fig. 7, and 8. Separate UAV TIR surveys are designated by a combination site location abbreviation (capital letters) and a sequence number (ordered from first to last survey conducted); H1 to H5 for Hardin Co., M1 to M7 for Morrow Co., RA1 to RA6 for Ross Co. A, RB1 to RB4 for Ross Co. B, and S1 to S6 for Seneca Co. Again, UAV TIR survey sets at Hardin Co.

Conclusions

There are both economic and environmental reasons for finding viable and nondestructive means for mapping agricultural subsurface drainage systems. Previous studies indicate that unmanned aerial vehicle (UAV) thermal infrared (TIR) imagery has drainage mapping potential. Knowing the best times of day to conduct UAV TIR surveys for drainage mapping is important for the successful employment of this technology within agricultural settings. In order to determine the best time of day for carrying

Authors Note

The use of equipment manufacturer and software developer names is for informational purposes only and does not imply endorsement by the authors or their organizations.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This research did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors. The authors wish to express their appreciation to all the land owners who provided field access to conduct UAV surveys.

References (41)

  • P. Aitken et al.

    Chapter 1 – Regulations

    Part 107 Drone Certificate Study Guide

    (2017)
  • B. Allred et al.

    Delineation of agricultural drainage pipe patterns using ground penetrating radar integrated with a Real-Time Kinematic Global Navigation Satellite System

    Agriculture

    (2018)
  • B. Allred et al.

    Overall results and key findings on the use of UAV visible-color, multispectral, and thermal infrared imagery to map agricultural drainage pipes

    Agric. Water Manag.

    (2020)
  • B.J. Allred

    A GPR agricultural drainage pipe detection case study: Effects of antenna orientation relative to drainage pipe directional trend

    J. Environ. Eng. Geophys.

    (2013)
  • B.J. Allred et al.

    Location of agricultural drainage pipes and assessment of agricultural drainage pipe conditions using ground penetrating radar

    J. Environ. Eng. Geophys.

    (2010)
  • B.J. Allred et al.

    Detection of buried agricultural drainage pipe with geophysical methods

    Appl. Eng. Agric.

    (2004)
  • B.J. Allred et al.

    Important considerations for locating buried agricultural drainage pipe using ground penetrating radar

    Appl. Eng. Agric.

    (2005)
  • AZoSensors, 2018. Different factors which impact thermal imaging quality....
  • T.L. Chow et al.

    Identification of subsurface drain locations with ground-penetrating radar

    Can. J. Soil Sci.

    (1989)
  • A. Dai et al.

    Diurnal and semidiurnal variations in global surface wind and divergence fields

    J. Geophys. Res.

    (1999)
  • Cited by (15)

    • Subsurface drainage outlet detection in ditches and streams with UAV thermal infrared imagery: Preliminary research

      2022, Agricultural Water Management
      Citation Excerpt :

      This project’s research hypothesis can be stated as follows: “Subsurface drainage system pipe outlets in ditches and streams of the Midwest U.S.A. can be successfully mapped with UAV TIR imagery.” This research used much the same unmanned aerial vehicle (UAV) equipment and data processing procedures as described by Allred et al., (2020, 2021). A senseFly SA (Cheseaux-sur-Lausanne, Switzerland) eBee Plus RTK/PPK fixed-wing drone, with S.O.D.A. and thermoMap cameras, was employed to carry out the UAV surveys for this study.

    View all citing articles on Scopus
    View full text