Linear Transects of Imagery Increase Crop Monitoring Efficiency Using Fixed‐Wing Unmanned Aircraft Systems

Abstract

R emote sensing with small unmanned aircraft systems (UAS) is advantageous for agriculture (Zhang and Kovacs, 2012; Maes and Steppe, 2019) and may be classified into three general categories: scouting for problems, monitoring crop growth, and planning management operations (Hunt and Daughtry, 2018). Typically, UAS images are acquired with large front-to-back and side-to-side overlap to create an ortho-rectified mosaic image of the entire field. The mosaic images are used as maps for planning scouting visits. However, it is unknown if ortho-rectified mosaics are necessary for crop monitoring or planning management operations. Monitoring crop growth over time provides an early warning to prevent yield loss (Barbedo, 2019; Franceschini et al., 2019). But few studies have compared costs and returns of UAS with sensor-based alternatives. For monitoring irrigation, UAS remote sensing had higher costs compared with a ground-based wireless sensor network (West and Kovacs, 2017). The cost of UAS is smaller than aircraft or satellite imagery for a 5-ha field but is much larger for a 50-ha field (Matese et al., 2015). Thus, monitoring production-scale farms with UAS may be cost prohibitive. Therefore, new methods are needed to reduce costs of using UAS remote sensing, particularly for crop monitoring. Previously, we proposed that UAS flight lines could be used to sample fields based on linear transects, where each image is analyzed as one sample point along the transect (Hunt et al., 2014; Hunt and Daughtry, 2018). In this study, we used a fixed-wing UAS to acquire imagery to demonstrate the potential of flight-line transects for crop monitoring. Methods Six 50-ha, center-pivot irrigated fields were selected in north-central Oregon, USA, at about 45.68° latitude,-119.89° longitude, and 215 m elevation above mean sea level. (Fig. 1A).