AqUavplant Dataset: High-Resolution Aquatic Plant Images from Drones
Author Information
Author(s): Istiak Md. Abrar, Khan Razib Hayat, Rony Jahid Hasan, Syeed M. M. Mahbubul, Ashrafuzzaman M., Karim Md. Rajaul, Hossain Md Shakhawat, Uddin Mohammad Faisal
Primary Institution: Independent University, Bangladesh
Hypothesis
The study aims to develop a comprehensive dataset for monitoring and mapping aquatic plant species using UAV technology.
Conclusion
The AqUavplant dataset consists of 197 high-resolution images of 31 aquatic plant species, which can aid in biodiversity monitoring and conservation efforts.
Supporting Evidence
- The dataset includes images of 31 aquatic plant species.
- Images were collected from nine different locations in Bangladesh.
- The drone used was the DJI Mavic 3 Pro, ensuring high-resolution image capture.
- Data annotation was performed for both binary and multiclass semantic segmentation.
- The dataset can be used for monitoring plant growth and biodiversity.
- Machine learning models can be developed using this dataset for automatic plant mapping.
- The study addresses limitations of traditional mapping methods.
Takeaway
This study created a collection of pictures of water plants taken by drones to help scientists understand and protect these plants better.
Methodology
Images were captured using a DJI Mavic 3 Pro drone, followed by data annotation for semantic segmentation.
Potential Biases
The disproportionate ratio of instances among species classes may lead to misclassification in minority classes.
Limitations
The dataset may have imbalances in species representation, which could affect model performance.
Participant Demographics
Images were collected from nine different sites in Bangladesh.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website