Logo des Repositoriums

Auflistung nach Schlagwort "plant detection"

1 - 2 von 2
  • Konferenzbeitrag
    Evaluating synthetic vs. real data generation for AI-based selective weeding
    (43. GIL-Jahrestagung, Resiliente Agri-Food-Systeme, 2023) Iqbal, Naeem; Bracke, Justus; Elmiger, Anton; Hameed, Hunaid; von Szadkowski, Kai
    Synthetic data has the potential to reduce the cost for ML training in agriculture but poses its own set of problems compared to real data acquisition. In this work, we present two methods of training data acquisition for the application of machine vision algorithms in the use case of selective weeding. Results from ML experiments suggest that current methods for generating synthetic data in the field of agriculture cannot fully replace real data but may greatly reduce the quantity of real data required for model training.
  • Konferenzbeitrag
    Field plant characterization method based on a multi-wavelength line profiling system
    (41. GIL-Jahrestagung, Informations- und Kommunikationstechnologie in kritischen Zeiten, 2021) Pamornnak, Burawich; Scholz, Christian; Nieberg, Dominik; Igelbrink, Matthias; Ruckelshausen, Arno
    Phenotyping of plant characteristics is essential for plant breeding. Especially the growth stages of plants during field emergence, described by parameters such as plant height and plant counting, are of interest. But large-scale manual phenotyping is very inconvenient due to the workload, the harsh weather conditions, and time-consumption. Therefore, an automated system is needed. This research describes a field plant characterization method implemented in a plot divider machine for rapeseeds. The method consists of a plant height estimation and a plant counting system. Based on a multi-wavelength line profiling (MWLP) sensor system, the 2D and 3D point cloud information from visible wavelengths to near-infrared (NIR) are automatically mapped without any need for a matching method. The plant characterization processes consist of two main steps, 1) plant detection, and 2) height estimation. These processes use the 2D NIR and 3D point cloud as the main features. The proposed method was demonstrated with highly accurate results in several rapeseeds, illustrating the potential of this method to become a basic tool for crop characterization in plant sciences