Stray Obster Survey (remote sensing)

The data comes from a nationwide survey of the orchards using photogrammetric aerial images with a spatial resolution of 40 cm, i.e. 6.25 points/m² (fly 2015-2018). Trees with a height between 3 m and 20 m are addressed as stray fruit trees according to a specific classification scheme and taking into account the other parameters. Not all trees were classified with the same trustworthiness. Shown are points of tree classes 3-5, where the probability that they are scattered fruit trees is comparatively high. Each point represents a single tree. The classification quality was determined on the basis of a validation. Here, the orchards were mapped on 13 areas in the terrain. In the classification, 79.4 % of the stray fruit trees mapped in the field were recorded, i.e. 20.6 % of the stray fruit trees, especially juvenile trees, could not be recognised as such using the classification algorithms or were excluded from the classification by the set parameter threshold values. For grades 3-5, 77.0 % of the trees were correctly classified as a stray fruit tree. 24.0 % of the trees recorded by remote sensing were incorrectly classified as a stray fruit tree. For more information on the survey and evaluation, see Borngräber et al. 2020 (BORNGRABER, S., A. KRISMANN & K. SCHMIEDER (2020): Determination of the clustered fruit stocks of Baden-Württemberg by automated remote sensing procedures. — Nature conservation and landscape care Baden-Württemberg, Volume 81 (Online-First)) No evaluation was carried out for 0.3 % of the land area (96 tiles with 1 km² each within the national border). The validity of the data is limited due to the methodology used and requires validation in the field. These are not results of mapping.