Correcting position of delayed on-the-go field measurements by optimizing nearest neighbor statistics
-
A. González Jiménez [1] ; Yakov Pachepsky [2] ; J.L. Gómez Flores [3] ; M. Ramos Rodríguez [3] ; Karl Vanderlinden [3]
-
[1] Instituto de Investigación y Formación Agraria y Pesquera
Instituto de Investigación y Formación Agraria y Pesquera
Sevilla, España
-
[2] Agricultural Research Center
Agricultural Research Center
Egipto
-
[3] Instituto de Investigación y Formación Agraria y Pesquera (Córdoba)
-
- Localización: Estudios en la zona no saturada del suelo. Vol. XV, ZNS'21, A Coruña, 9-10 noviembre 2021: trabajos presentados en las XV Jornadas de Investigación en la Zona no Saturada del Suelo / Javier Samper (ed. lit.), Antonio Paz González (ed. lit.), Jorge Dafonte Dafonte (ed. lit.), Eva Vidal-Vázquez (ed. lit.), 2021, ISBN 978-84-9749-821-0, págs. 31-36
- Idioma: inglés
- Enlaces
-
Resumen
On-the-go field measurements of soil andplant characteristics, including yield, are commonplace incurrent Precision Agriculture applications. Yet, suchmeasurements can be affected by positional inaccuraciesthat result from equipment configuration and delays in thedata transmission, reception or logging. The resulting timeand position lags cause a misfit between the measurementsand their attributed GPS position. To compensate for thiseffect a simple coordinate translation along themeasurement direction is proposed, depending on the localvelocity and a specific time lag, which is estimated byminimizing the average absolute difference between thenearest neighbors. The correction procedure is evaluatedusing electromagnetic induction data with different spatialconfigurations and by comparing variograms for correctedand raw data. Best results are obtained when overlappingmeasurements are available, while the worst results arefound when no overlapping measurements exist. Furtherimprovements in the nearest neighbor search algorithm arediscussed. The results are valid beyond motorized soilsensing applications.
