home.social

#tillagedetection — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #tillagedetection, aggregated by home.social.

fetched live
  1. 💧🌏 Greg Cocks @[email protected] ·

    Optical, Radar, And Hybrid Indices To Detect Farming Practices In Europe
    --
    doi.org/10.1016/j.rse.2026.115 <-- shared paper
    --
    “HIGHLIGHTS:
    • [they] compare[d] Sentinel-1 and Sentinel-2 time series to detect farming practices.
    • HyBRIS index is introduced, temporally weighting BSI and VH/VV into a daily index.
    • Time-series minima and maxima are used to predict sowing, harvest, and tillage.
    • Validation is performed across several years, crop types, and European locations.
    • Phenology detection is improved compared to HRL-Cropland.
    ABSTRACT: Arable farming practices dictate both crop cycles and soil dynamics, and are central to agriculture's environmental impact and its mitigation. Sowing and harvesting mark the beginning and end of the growing season, while tillage modifies soil structure during the dormant period. Although well-established methods exist for delineating the growing season using phenology and optical data, the detection of farming practices, particularly tillage, remains underexplored. This study investigates the strengths of radar and optical data to retrieve sowing, harvest, and tillage dates at the field level, and proposes a novel Hybrid Bare Soil Radar Index (HyBRIS). Based on Sentinel-1 and Sentinel-2, HyBRIS merges optical and radar data into a single index using a temporally weighted mean. Local minima and maxima of the time series are used to detect farming practices across European sites. Validation is carried out against a reference dataset comprising 238 fields in 11 EU countries, including 462 sowing, 374 harvest, and 388 tillage events covering more than 40 crop types over 8 years. Compared to the Copernicus High Resolution Layer Croplands product (HRL-Cropland), the proposed method based on HyBRIS time series improved sowing and harvest dates detection (MAE 26 and 23 days, respectively). Additionally, this method enabled tillage dates estimation during dormant periods (MAE = 28 days), but tended to overestimate the number of tillage events (producer's accuracy = 97%, user's accuracy = 70%). Incorporating soil moisture data is advised for reducing false positives. The results highlight the potential of optical, radar, and hybrid indices for monitoring agricultural management and supporting environmental stewardship…”
    #Sowing #Harvest #tillage #tillagedetection #cropland #CroplandManagement #remotesensing #earthobservation #sentinel #Copernicus #cropland #satellite #optical #radar #sensor #landuse #landcover #landsurface #phenology #agricultural #monitoring #GIS #spatial #mapping #spatialanalysis #spatiotemporal #arable #farming #agriculture #soil #substrate #environment #sustainability #environmentalstewardship #growingseason #Europe #region #model #modeling

    #sowing #harvest #tillage #tillagedetection #cropland #croplandmanagement
  2. 💧🌏 Greg Cocks @GregCocks ·

    Optical, Radar, And Hybrid Indices To Detect Farming Practices In Europe
    --
    doi.org/10.1016/j.rse.2026.115 <-- shared paper
    --
    “HIGHLIGHTS:
    • [they] compare[d] Sentinel-1 and Sentinel-2 time series to detect farming practices.
    • HyBRIS index is introduced, temporally weighting BSI and VH/VV into a daily index.
    • Time-series minima and maxima are used to predict sowing, harvest, and tillage.
    • Validation is performed across several years, crop types, and European locations.
    • Phenology detection is improved compared to HRL-Cropland.
    ABSTRACT: Arable farming practices dictate both crop cycles and soil dynamics, and are central to agriculture's environmental impact and its mitigation. Sowing and harvesting mark the beginning and end of the growing season, while tillage modifies soil structure during the dormant period. Although well-established methods exist for delineating the growing season using phenology and optical data, the detection of farming practices, particularly tillage, remains underexplored. This study investigates the strengths of radar and optical data to retrieve sowing, harvest, and tillage dates at the field level, and proposes a novel Hybrid Bare Soil Radar Index (HyBRIS). Based on Sentinel-1 and Sentinel-2, HyBRIS merges optical and radar data into a single index using a temporally weighted mean. Local minima and maxima of the time series are used to detect farming practices across European sites. Validation is carried out against a reference dataset comprising 238 fields in 11 EU countries, including 462 sowing, 374 harvest, and 388 tillage events covering more than 40 crop types over 8 years. Compared to the Copernicus High Resolution Layer Croplands product (HRL-Cropland), the proposed method based on HyBRIS time series improved sowing and harvest dates detection (MAE 26 and 23 days, respectively). Additionally, this method enabled tillage dates estimation during dormant periods (MAE = 28 days), but tended to overestimate the number of tillage events (producer's accuracy = 97%, user's accuracy = 70%). Incorporating soil moisture data is advised for reducing false positives. The results highlight the potential of optical, radar, and hybrid indices for monitoring agricultural management and supporting environmental stewardship…”

    #sowing #harvest #tillage #tillagedetection #cropland #croplandmanagement