Publicação
A segmentation approach for object detection on highly dynamic aquatic environments
| Resumo: | The majority of segmentation methods applied to outdoor environments are not suitable to highly dynamic backgrounds, like in an outdoor swimming pool. This paper presents a new method for image segmentation of objects with a background with random noise generated by water oscillation. The key feature of this segmentation method is based on processing hue component of color in HSV color space on a pre-determined image area. This component is not affected by reflections, splashes, random water movements and overall scene light changes on the water surface. Foreground is extracted based on combination of temporal background estimation with spatial correction and inter-frame subtraction. The algorithm was applied and tested in a real situation, specifically on an outdoor domestic pool, showing good robustness and reliability with less computational load. |
|---|---|
| Autores principais: | Peixoto, Nuno |
| Outros Autores: | Cardoso, Nuno; Cabral, Jorge; Tavares, Adriano; Mendes, A. José |
| Assunto: | Image segmentation Automated video surveillance |
| Ano: | 2009 |
| País: | Portugal |
| Tipo de documento: | comunicação em conferência |
| Tipo de acesso: | acesso restrito |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | The majority of segmentation methods applied to outdoor environments are not suitable to highly dynamic backgrounds, like in an outdoor swimming pool. This paper presents a new method for image segmentation of objects with a background with random noise generated by water oscillation. The key feature of this segmentation method is based on processing hue component of color in HSV color space on a pre-determined image area. This component is not affected by reflections, splashes, random water movements and overall scene light changes on the water surface. Foreground is extracted based on combination of temporal background estimation with spatial correction and inter-frame subtraction. The algorithm was applied and tested in a real situation, specifically on an outdoor domestic pool, showing good robustness and reliability with less computational load. |
|---|