基于阈值法和区域生长法的图像分割算法研究

由于图像阈值法的直观性和易于实现的性质,使它在图像分割领域中处在中心地位,但是在阈值分割后不可避免地会出现与目标灰度接近的背景区域,从而产生了噪声干扰。由于图像分割技术的不断完善,区域生长法比较完美地解决了噪声干扰的问题,但是随之也会造成欠分割和过分割的现象,同样使得分割效果不理想。因此,本文采用阈值分割法和区域生长法相结合的方法,来尽可能避免目标本身灰度不均匀造成的欠分割现象和有效防止目标边界模糊造成的过分割现象。

Research on image segmentation based on threshold value method and regional growth method

Image threshold value method play an important role in the image segmentation field because of its intuitiveness and realizability. However there will appear background region which is similar to object region after threshold segmentation inevitably,and then produced noise interference.due to the continuous improvement of the technology of image segmentation,regional growth method perfectly solved the problem of noise interference, but it also can cause owe segmentation and over-segmentation phenomena, and also make segmentation effect is not ideal.therefore, we adopt the method of combining threshold segmentation method and the method of regional growth method,and to avoid under-segmentation phenomenon As far as possible caused by target area gray not even and to prevent over-segmentation phenomena caused by target fuzzy boundaries effectively.