Effects of Foreground Augmentations in Synthetic Training Data on the Use of UAVs for Weed Detection.
Simon HallöstaMats Ingemar PetterssonMattias DahlPublished in: NLDL (2024)
Keyphrases
- training data
- object detection
- bounding box
- learning algorithm
- classification accuracy
- automatic detection
- detection method
- detection accuracy
- dynamic environments
- training set
- domain knowledge
- prior knowledge
- decision trees
- unmanned aerial vehicles
- data sets
- false positives
- false alarms
- background subtraction
- training dataset
- generalization error
- control algorithm
- mission planning
- noisy data
- detection rate
- closed form
- training examples
- detection algorithm
- image sequences
- real world
- neural network