【摘要】 图像分割是图像识别中的主要问题 ,同时它又是一个经典难题 ;声呐图像恶劣的品质使图像的分割变得更为困难 .属性直方图概念是直方图概念的推广 ;Otsu阈值化方法是一种性能良好的方法 .本文将属性直方图概念应用于Otsu法中 ,提出了一种基于属性直方图的图像分割方法 .文中阐述了这种方法的原理 ,并以声呐图像分割作为实例说明了这种方法的可行性 .更多还原

【Abstract】 Image segmentation is an important problem on pattern recognition and one of the classical difficult problems,and sonar image segmentation is still an important problem because of bad quality. The concept of the bounded histogram is developed from the concept of histogram. Otsu’s method is a good thresholding method. The bounded histogram is applied to Otsu’s method and an image segmentation method based upon the bounded histogram,is proposed.Its principle is expatiated on and its feasibility is demonstrated by an example of sonar image segmentation.更多还原