Underwater image enhancement based on fusion of intensity transformation techniques

• Laura A. Martinho UFAM
• Felipe G. Oliveira UFAM
• João M. B. Cavalcanti UFAM
• José L. S. Pio UFAM

Resumo

The exploration of underwater environments has increased in recent years. The use of underwater robotic vehicles for search and rescue, monitoring and sea life maintenance has been the subject of research involving the academic and industrial segments. Under this light, the enhancement of underwater images is a paramount stage for the execution of the tasks above. aintIn this paper, we propose an underwater image enhancement approach to improve the quality of underwater images through the fusion of image transformation methods. To validate the proposed methodology, experiments using real-world underwater images were carried out. From qualitative and quantitative analysis we demonstrate the accuracy and effectiveness of the proposed approach. For the Peak Signal-to-Noise Ratio (PSNR) metric the obtained result is 23.185. Meanwhile, for the Structural Similarity Index (SSIM) metric the obtained result is 0.859, outperforming state-of-the-art underwater image restoration techniques.

Palavras-chave: Measurement, Image segmentation, PSNR, Statistical analysis, Service robots, Maintenance engineering, Indexes