A New De-Noising Method for Ground Penetrating Radar Signal

Ma, Yong-Min; Jong, Yong-Gwang; Liu, Yang

doi:10.1088/1742-6596/1802/2/022002

Cited by 1 publication

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An additional alternative for GPR feature extraction is the use of wavelet methodologies. Wavelet analysis in GPR has been largely applied as a de-noising technique [24,25]; wavelets have also been applied in a diagnostic capacity in civil engineering [26] as well as in evaluation of soil moisture [27]. The fibrous and rhizomatous biomass assessed in this current study are characterized by a lack of a single large, discrete target object and are distributed among three depth layers.…”

Section: Introductionmentioning

confidence: 99%

Wavelet Analysis of GPR Data for Belowground Mass Assessment of Sorghum Hybrid for Soil Carbon Sequestration

Wolfe,

Dobreva,

Ruiz-Guzman

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

Among many agricultural practices proposed to cut carbon emissions in the next 30 years is the deposition of carbon in soils as plant matter. Adding rooting traits as part of a sequestration strategy would result in significantly increased carbon sequestration. Integrating these traits into production agriculture requires a belowground phenotyping method compatible with high-throughput breeding (i.e., rapid, inexpensive, reliable, and non-destructive). However, methods that fulfill these criteria currently do not exist. We hypothesized that ground-penetrating radar (GPR) could fill this need as a phenotypic selection tool. In this study, we employed a prototype GPR antenna array to scan and discriminate the root and rhizome mass of the perennial sorghum hybrid PSH09TX15. B-scan level time/discrete frequency analyses using continuous wavelet transform were utilized to extract features of interest that could be correlated to the biomass of the subsurface roots and rhizome. Time frequency analysis yielded strong correlations between radar features and belowground biomass (max R −0.91 for roots and −0.78 rhizomes, respectively) These results demonstrate that continued refinement of GPR data analysis workflows should yield an applicable phenotyping tool for breeding efforts in contexts where selection is otherwise impractical.

show abstract