Fusion of Reflected GPS Signals With Multispectral Imagery to Estimate Soil Moisture at Subfield Scale From Small UAS Platforms

Abstract

This study proposes a low-cost and "proof-of-1 concept" methodology to obtain high spatial resolution soil 2 moisture (SM) via processing reflected Global Positioning System 3 (GPS) and a multispectral camera data acquired by small 4 Unmanned Aircraft System (UAS) platforms. An SM estimation 5 model is developed using a random forest (RF) machine-learning 6 (ML) algorithm by combining features obtained from reflected 7 GPS signals (collected by smartphones and commercial off the 8 shelf receivers) in conjunction with ancillary vegetation indices 9 from the multispectral camera data. The proposed ML algorithm 10 uses in-situ SM measurements acquired via SM probes as labels. 11 A preliminary field experiment was conducted on 210 m by 110 m 12 (2.31 ha) crop fields (corn and cotton) in 2020 (from January 13 to November, including crop planting through senescence time 14 period) at Mississippi State University (MSU)'s the heavily 15 instrumented North Farm to acquire data needed for the ML 16 model to train and test. Our results showed that both fields 17 could be covered by GPS reflectometry measurements with about 18 13 minutes of flight time at a 15-m altitude, and SM can be 19 mapped with 5m × 5m spatial resolution (corresponding to 20 the elongated first Fresnel zone). The model is trained with 21 and validated against eight in-situ SM station datasets via 10-22 fold and leave-one-probe-out cross-validation techniques. Overall 23 root-mean-square errors (RMSE) of 0.013 m 3 m −3 volumetric SM 24 and R-value of 0.95 [-] are obtained for 10-fold cross-validation. 25 The proposed model reached an RMSE of 0.033 m 3 m −3 and an 26 R-value of 0.5 [-] in leave-one-probe-out cross-validation. While 27 having limited data, the results indicate that high resolution SM 28 measurement can be achieved with a low-cost GPS reflectometry 29 system onboard a small UAS platform for use in precision 30 agriculture (PA) applications.31

GNSS-R snow depth retrieval algorithm based on PSO-LSTM

“…In the field of remote sensing, GNSS-R technology has received increasing attention, and many scientists are continuously researching it. GNSS-R technology has been widely used to monitor sea level height [4][5][6], snow depth [7][8][9], soil moisture [10][11][12], and sea ice detection [13,14], etc. In the field of snow depth retrieval technology, Larson et al [15] used global positioning system (GPS) receivers and traditional methods to monitor heavy snowfall, respectively, and the experimental results showed that the snow depth could be effectively monitored using GPS receivers, and the monitoring results were very similar to the classical methods.…”

Preliminary inquiry on the linear relationship between the height of the station and the ground height error retrieved by GNSS-IR with low-cost smart electronic equipment

“…Due to its outstanding height measurement performance, GNSS-IR technology is often used to detect snow depth (Larson et al 2009, Tabibi et al 2017, Zheng et al 2022a, sea level height (Jin et al 2011, Larson et al 2013, Löfgren et al 2014, Wang et al 2019, Zheng et al 2021, soil moisture (Chew et al 2014, Senyurek et al 2022, and sea ice thickness (Strandberg et al 2017 and so on. However, only a few scholars have studied the height measurement performance of low-cost equipment, let alone the impact of different station heights on measurement performance.…”