Water Hyacinth Segmentation for Aquatic Weed Elimination in Thailand

摘要

The water hyacinth is a type of aquatic weed that could harm the ecosystem. As it is tolerant to extreme environment, it has spread a over very large area of Thailand. It has become one of the main obstacles in dam irrigation management. Therefore, it needs to be eradicated. In this paper, we propose an algorithm to determine the area in the river or dam that is covered with water hyacinth. The algorithm was tested by using multiple perspective images. The optimal color space was identified to facilitate the optimum segmentation of the water hyacinth. We found that the use of the proposed algorithm with the HSL color space results in 95.65% accuracy as compared to roughly 80% accuracy when using other color spaces.