Publications by authors named "Raymond Ajekwe"
Plant Methods
August 2023
Article Synopsis
- Efficient and site-specific weed management is essential in agriculture, and using drone images with machine learning can improve weed assessment, but image quality issues like motion blur can hinder this process.
- The study introduces DeBlurWeedSeg, a model that integrates deblurring and segmentation to effectively identify weeds and crops in motion-blurred images, utilizing a new dataset of matched sharp and blurred images collected from drones.
- DeBlurWeedSeg significantly outperforms traditional segmentation methods that lack deblurring capabilities, enhancing accuracy in weed identification, which is crucial for improving agricultural practices like robotic weed removal.
