An interval-valued fuzzy entropy approach to object detection with dynamic thresholding for underwater images

作者：Deng, Tingquan^*; Wang, Zhanjiang; Wang, Peipei

来源：Journal of Harbin Engineering University, 2011, 32(2): 246-250.

DOI：10.3969/j.issn.1006-7043.2011.02.019

• Canny edge detection
• Dynamic thresholding
• Fuzzy entropy
• Image object detection
• Interval-valued
• Interval-valued fuzzy sets
• Optimal entropy
• Thresholding segmentation

In view of shortcomings with existing versions, a new characterization for interval-valued fuzzy entropy sets was proposed and applied to image segmentation. An image was represented by an interval-valued fuzzy set, and an initial threshold for image segmentation was determined by the principle of optimal entropy. In combining the initial threshold with the result of classical Canny edge detection, a novel dynamic thresholding technique for image segmentation was proposed to meet the aim of underwater image object detection in complex circumstances. Simulated examples show the rationality and practicality of the new technique for image segmentation and object detection.

• 出版日期2011
• 单位哈尔滨工程大学